KANTIAN RATIONALISM

Kantian rationalism represents an important and influential approach to the justification of moral principles. For Kant, the source of morality is, and must be, our own practical reason. Moral principles are derived from reason and these principles express our reason-governed free agency. As Christine Korsgaard has perhaps best explained, the source of morality cannot be an external law-giver (as suggested by Divine Command Theories and in a different way by Hobbesian Theories), for then the actual rightness of the action could not function as the agent's motive as the concept of duty requires. Alternatively, the source of morality cannot be a natural sentiment we find in ourselves (as Moral Sense Theories suggest), for then moral actions would not be a reflection of the person's agency and thus could not be attributable to the person as required by moral responsibility.Footnote ² The concept of morality presupposes that our practical reason can serve as a legislative will that is the source of our principles of action (and that, with time, also shapes and transforms our perfectible human nature).Footnote ³ At the core of the Kantian rationalist conception of morality is the idea that deliberative rationality is constituted by internal standards that limit and shape the outcome of rational deliberation. The justification of actions presupposes that the principle, purporting to justify action, meets some standard of justification. These standards are not external principles or values; they are instead internal or intrinsic to deliberative rationality itself.

The Kantian universalizability procedure is the paradigm case of an internal standard that constrains the outcome of (moral) deliberation. Kant argues that moral principles are categorical imperatives, which must have the form of a universal and unconditional law for all rational agents. He concludes that we can test the morality of an action by asking whether the agent could will to do the action and could at the same time will that the principle of action hold as a universal law of nature. If my proposed principle (or maxim) does not pass this procedural test it cannot be a moral principle. This test is supposed to be internal and intrinsic to the essential nature of a moral principle. Of course, even many Kantians recognize that the universalizability procedure alone does not provide a sufficient conception of the supreme principle of morality.Footnote ⁴ Kant's conception of human dignity, that is the special value of rational nature or humanity as an end-in-itself, is also necessary to give life and substance to Kantian ethics.

Kant argues that rational nature sets itself apart from the rest of nature because it involves the capacity to set oneself an end and to act on principle. Kant also argues that rational nature is an end-in-itself and that the essence of morality is treating oneself and others always as an end and never as a mere means to one's own subjective ends. If we do not treat rational nature as an end-in-itself, Kant argues, we cannot justify our actions to ourselves or to others. All actions that fail to treat rational nature as an end-in-itself are unjustifiable and thus wrong. Indeed, Kant thinks that if we do not treat rational nature as an end, then, in so acting, we undermine the basis and source of the value of our own ends. This step in Kant's argument is controversial but it is a core principle of Kantian ethics that rational nature is an end-in-itself. However one reconstructs the Kantian argument, the important point here is that the end-in-itself must be an end for all rational beings, and thus Kant argues it must be something that all rational agents share as an end. The only thing that all rational beings necessarily share is rational nature itself – the capacity to set oneself ends and act on principle. Although self-reflection is as much a process of discovery as choice, it is also the case that our ends are not fixed. This capacity to critically evaluate our ends and reflectively endorse a conception of the good is a distinctive capacity of persons. Rational nature sets itself apart from the rest of nature because of its capacity to rationally revise its ends and conception of the good.Footnote ⁵

According to Kant, only the ‘autonomy of the will’ can function as the authoritative source of justifying reasons and thus also of moral reasons (understood as categorical imperatives). This conclusion follows from Kant's arguments against alternative conceptions of the source and authority of morality. Once we understand the problems with divine command theories, moral sense theories, substantive moral realism and dogmatic rational intuitionism, then we also see that the autonomy of the will must be the source of morality and that all other conceptions of ethics are forms of heteronomy.Footnote ⁶ Kant also argues that our deliberative rational capacities provide an internal, intrinsic, objective constraint on the will. Freedom is not lawless but a form of rational causality, and the rational constraint is that we treat rational nature in ourselves and others as an end-in-itself. In forming and revising my ends, I thus have a duty to recognize, in my principles and actions, my own rational nature as the source of justifying reasons. However, since it is rational nature itself that sets itself apart and must be treated as an end-in-itself, Kant argues, I must treat rational nature in myself and in others as a source of reasons for action. One interpretation of the idea here is that if I do not respect others, then my principle of action permits others to not respect me, and it thus undermines and expresses disrespect for myself as the source of the reason-giving value to my ends. So principles of actions that do not respect rational nature cannot be justified. Respect for the end-setting capacity of others, and also the ends that they have set, is thus a constraint on the ends and principles that I can consistently endorse. Kant thus concludes that all persons must treat rational nature both in themselves and in others as an end-in-itself.

In one sense rational nature, and reason-governed freedom, is itself like a substantive value. The value of autonomy does give substance and content to Kant's other more formal formulations of the categorical imperative. In many contexts there is no harm in thinking of autonomy as a substantive value, for it functions like a substantive value. It is not, however, a substantive value in a metaphysical realist sense. It is instead an objective practical value; that is, its value is a necessary presupposition of the deliberative perspective of free rational agency. The idea of a practical value, in this sense, is analogous to the Kantian idea of practical freedom, and it is also related to the practical contradiction interpretation of Kant's first formulation of the categorical imperative.

First, freedom itself is only a practical idea for Kant. It is not a metaphysical idea. Kant argues that it is a necessary presupposition of deliberative action that we ‘act under the idea of freedom’. We will not enter into the complex questions raised by a full investigation of Kant's conception of freedom. Suffice it to say that the current Kantian interpretation, with which I agree, insists that Kant's is a practical conception of freedom, which is distinct from a metaphysical conception of freedom. Since practical freedom does not imply a metaphysical incompatibilism, the value of autonomy does not necessarily involve a substantive metaphysical conception of freedom. If human dignity is founded on autonomy, and autonomy is a form of practical freedom, then a Kantian conception of dignity should not assume that human dignity implies a special metaphysical status. The special value of autonomy simply is not a substantive metaphysical value and it thus does not imply, and indeed it is inconsistent with, the recognition of any substantive value that transcends the value of deliberative freedom itself.Footnote ⁷

Second, the idea that the value of autonomy is to be understood as a practical value also fits perfectly with a plausible interpretation of the universalizability test found in Kant's formula of universal law. Kant argued that we could test a proposed principle of action (which he calls a maxim) by checking to see if we can will, at the same time, the proposed principle and its universalization. If the attempted universalization results in a contradiction in the conception of the universalized maxim itself or in the agent's will in universalizing the maxim, then the proposed principle cannot justify the action in question. Korsgaard has argued, convincingly I think, that the contradiction here is not a logical or teleological contradiction; it is a ‘practical contradiction’. The maxim when universalized defeats the agent's own fundamental purpose in acting and thus the principle cannot serve as a justification of the action. The formula of universal law still has many problems, and it seems pretty clear that the universalizability test alone is not the supreme principle of morality. The idea of a practical contradiction, however, is precisely the right conception for understanding the idea of autonomy as a practical value.Footnote ⁸ Kant argues that persons must value their own autonomy as the condition that makes justified action possible. As the source of justifying reasons, autonomy itself is a value that must be presupposed by practical reason and as such it thus limits what can even count as a justifying reason.

Consciousness of our capacity for autonomy, and thus of the moral law, is the basis of our recognition of the dignity of humanity. Respect for persons is respect for the moral law that is within each person, for it is the capacity to will moral law that renders us free and responsible moral agents. Autonomy is the basis of our elevated status and it is thus the ground of human dignity. Although the basis of human dignity is the capacity to will moral law, that is, unconditional and universal practical principles, these principles must treat rational nature as an end-in-itself. So we respect the dignity of persons by respecting the moral law in each person, and we respect the moral law in each person by treating rational nature as an end-in-itself.

Kant's conception of the dignity of humanity is thus essentially linked to the idea that rational nature must serve as an end-in-itself. The unconditional value of rational nature, and thus the special status and dignity of humanity, is itself justified by the intrinsic constraints of deliberative rationality. There is no appeal to any substantive value external to rational nature itself. The special practical value of rational nature follows directly from its distinctive capacity to set itself ends and act on principle.

Now, Kant famously contrasts dignity with price, and argues that persons with dignity have an unconditional value that has no price.Footnote ⁹ Persons are not things or commodities, and thus they have a special dignity. The ends that are based on inclinations have a relative value and they are interchangeable in the currency of inclination. Indeed, their value is set by the strength of inclination (or preferences) and thus they have a mere ‘market price’ in the currency of commodities. Indeed, preferences have practical value and give rise to practical reasons because persons have a prior value. If persons did not have value in themselves, their preferences would not have value either. The value of a person's preferences is derivative and secondary to the value of persons themselves.Footnote ¹⁰ The distinction between dignity and price must itself be based on the preceding analysis of the idea that persons are to be treated as ends and not mere means to my subjective ends. We cannot here simply introduce a new metaphysical conception of value that is unconnected with all that has come before.

One straightforward reading of the dignity–price distinction that is consistent with the previous arguments is that the value of rational nature is not reducible to the maximal satisfaction of inclinations, preference satisfaction, or even to the maximal realization of the ends (or conceptions of the good) that are reflectively endorsed by each rational agent. This reading is consistent with the idea that there are two kinds of value, the value of rational nature and the value of the realization of ends, and that the former is prior to and not interchangeable with the latter. Of course, Kantians often assume that the idea of dignity implies that the value of persons is non-comparable, and non-substitutable, in a way that is only captured by deontological constraints. This, however, requires additional argument (like Kamm's inviolability argument that we will discuss below); it cannot just simply be assumed. From what has been shown so far, we respect dignity by acting on universalizable principles and by treating persons as ends-in-themselves. So if there is a form of consequentialism that is universalizable and treats persons as ends-in-themselves, then it also respects human dignity.

CONSEQUENTIALISM AND THE DIGNITY OF PERSONS

Kant's conception of the universal law formula does not exclude a consequentialist normative principle. I have argued for this in the past, and Hare and Kagan have also defended a form of ‘consequentialism for Kantians’.Footnote ¹¹ Although I will not recite these arguments here, it is worth noting that Korsgaard, O'Neill and others present no argument at all for the conclusion that consequentialism fails the universalizability test. Of course, they do argue against some particular maxims of deception or coercion, but they do not adequately consider the case of a maxim of deception or coercion for the sake of maximizing the good, and they do not even discuss a fundamental consequentialist maxim. If my fundamental maxim of action is to act in ways that tend to maximize the good (however this is construed), then how does it defeat my purpose (and result in a practical contradiction), if everyone were to adopt and act on the same fundamental maxim? Far from defeating the purpose of my maxim, if all persons were committed to promoting the good, as a general matter, it would enhance and advance the end in question.Footnote ¹² Of course, a sophisticated consequentialist emphasizes that our specific actions and maxims, our decision procedures, would not always aim at maximizing the good. A sophisticated consequentialist would adopt secondary principles and rules, and inculcate virtues that work in coordination with others to promote the overall good. But the fundamental principle (the maxim of ends) that determines the shape of our intuitive day-to-day morality would still be the consequentialist principle to maximize the good.Footnote ¹³

Of course, it is the second formula of the categorical imperative (the imperative to treat rational nature as an end-in-itself), and not the first universalizability formula, that provides the more direct challenge to consequentialism. In particular, the formula of the end-in-itself is supposed to provide an alternative conception of value and its place in normative theory and, as we shall see below, it is this formula that is supposed to provide the essential link to contractualism. Before we consider the contractualist interpretation, however, there is a straightforward and direct consequentialist incorporation of Kantian value considerations that we must first take into account.

On previous occasions, I have argued for a consequentialist interpretation of the Kantian imperative to treat persons as ends and thereby respect the dignity of humanity. Kantian consequentialism has two basic normative principles. The first principle is a requirement to promote the flourishing of rational agency. The details of what this involves are open to interpretation. We do not need to settle these details prior to evaluating the deontological alternatives to this direct consequentialist conception of respect for dignity. Nonetheless, it would seem that a capabilities conception of the good, as developed by Sen and Nussbaum, which identifies the good with the capability to achieve distinctively human functioning, is a promising approach here.Footnote ¹⁴ The first tier of value would involve the basic goods necessary to sustain a human life that retains the capabilities for agency. It would also include the goods necessary for the development and exercise of our capabilities for agency, including the capability to rationally revise our ends and reject established social norms. In addition, respect for ‘finite’ rational beings clearly includes protecting the basic security of persons and their claim to the basic subsistence goods necessary for an adequate life.

The second tier of value is focused on the social conditions and external goods that are necessary for an agent to effectively advance their own conception of the good. Given the relationship between agents and their ends, the goal here is not for me to advance the ends of others for them, but for them to be able to promote their own ends. This is a modification of Kant's claim that we should make other's ends our own. Instead we recognize the value of other's ends by promoting the circumstances for agents to effectively pursue their own conception of the good. This two-tiered approach recognizes the value of preference-satisfaction, but it emphasizes the priority of the agent's responsibility to advance his or her own conception of the good. It thus subordinates the good of preference-satisfaction to the value of persons as reflective agents. If preferences are to have reason-giving value for others, individual preferences must reflect the equal value and dignity of all persons. The value of persons is prior to the value of preferences.

The special status of rational nature is reflected in the priority of the first tier over the second. More specifically, a person's rational nature cannot be violated to promote the happiness or satisfaction of other persons. In this sense, respect for the flourishing of rational nature will involve respect for basic rights; that is, a claim that others should promote and protect access to the basic goods that each person needs for the flourishing of rational nature. Kantian consequentialism includes a constraint on violating these basic rights of persons in order to advance the contingent, subjective preferences of others. This is not, however, a classic deontological constraint because (in principle) it does not prohibit infringing the rights of some to better secure the rights of others. It does not prohibit minimizing rights violations.

Although Kantian consequentialism does not include classic deontological constraints, the first principle does constrain the second (we cannot sacrifice rational nature to maximize happiness), and it thus captures some of the intuitive appeal of constraints. This type of two-tiered consequentialist normative theory recognizes that rational nature has special value that is distinct from, and not comparable to, happiness that is understood as the satisfaction of ends. In this sense, it reflects a clear recognition of the special value and dignity of persons.Footnote ¹⁵ Kantian consequentialism thus takes a direct approach to the special status and dignity of persons. We respect the special status and dignity of rational nature by promoting the flourishing of rational nature.

With this simple sketch in hand, we leave aside doubts and questions about the basis of the priority of rational nature and also the more precise specification of these principles.Footnote ¹⁶ Our question is whether or not the Kantian argument for the special status of rational nature, and thus the dignity of humanity, provides a rationale for deontological constraints.

CONTRACTUALISM AND SUBSTANTIVE VALUES

We all know that Kantians are not supposed to be consequentialists. Kantian ethics takes many forms, but consequentialism is not supposed to be one of them. Kantians instead turn to contractualism to provide a deliberative procedure that treats persons as ends-in-themselves. Contractualist models do seem to provide a plausible proceduralist interpretation of how one shows respect for the dignity of persons. The basic idea behind the contractualist procedural conception is very straightforward. These theories link respect for persons with some form of consent condition being met. The consent condition itself is an issue of discussion and variation among procedural approaches. Actual consent, presumed consent, possible consent, hypothetical consent and even more complicated consent conditions have been developed. The general point that I wish to make against contractualist conceptions, however, does not depend on the details of the consent condition. We will focus on Scanlon's version of contractualism, but the important point for now is that, in order to respect the dignity of persons, the appropriate kind of consent condition must be met. Of course, the philosophical basis of the connection between consent and respect for dignity is found in the Kantian idea that autonomy is the ground of the dignity of humanity. Since autonomy is the basis of our special dignity, contractualists argue, we respect dignity by respecting autonomy, and we respect autonomy by securing the appropriate consent.Footnote ¹⁷

It is the capacity to set ourselves ends, in accordance with reason, which sets rational nature apart from the rest of nature, and is thus the basis of the special status and dignity of humanity. Respect for this elevated status and dignity involves the mutual recognition of each person's capacity to set ends in accordance with reason. The Kantian contractualist thus tries to construct a fair choice procedure, or set of consent conditions, that gives due regard to the equal status of each person as a being that can set its own ends. If the procedure reflects a mutual recognition of a common dignity, then its outcome is a reasonable constraint on the individual choices of each person. The particular constraints of morality are thus themselves constructed by us as an expression of our common dignity. Since the procedure is supposed to generate moral principles, there is indeed an objective fact of the matter about what is right and wrong. This ‘moral fact’, however, is not an external fact about the world independent from us. It is a fact about the outcome of our own reason-governed legislative will; it is what we recognize as reasonable principles for governing our lives. The resulting objectivity of morals is best described as ‘procedural realism’, since it does not involve a commitment to ‘metaphysical realism’. There are no moral facts ‘out there’ that are independent from the internal constraints of practical reason.Footnote ¹⁸

This type of contractualist approach thus appears to be consistent with Kantian rationalism. Contractualism, however, is a more indirect approach in that respect for persons is mediated by a consent condition, which is the source of moral principles. We must see how this indirect approach is able to lead to a different result from the direct consequentialist interpretation, and whether it can generate this result without appealing to substantive values or principles other than the value of rational nature itself.

For the Kantian consequentialist, rational nature has a special value that must be given due recognition in deliberation and action. The consequentialist assumes that the best way to honor and respect rational nature is to directly promote the conditions conducive to the flourishing of rational nature. The contractualist objects that this misunderstands the Kantian conception of the value of rational nature.Footnote ¹⁹ The contractualist argues that we honor and respect the value of rational nature as a source of reasons by insuring that our reasons for action are justifiable to others. If others consent to, or hypothetically would consent to, or cannot reasonably reject, my principle of action, then my conduct is justifiable and gives due regard to the deliberative perspective of others. Morality does not require that we promote the value of rational nature; it requires that we recognize the equal status of others by making sure that our principles of action are justifiable to others. Of course, consequentialists also defend their theory as an account of appropriate impartial regard for the interest of all. Many deontologists maintain, however, that contractualist models better capture the idea of equal status of persons. There are many versions and variations of the contractualist approach. Although we will focus on Scanlon's recent and influential account of contractualism, the problem with his account is a general problem faced by any contractualist approach.

Scanlon's theory is in many ways a development of Rawls’ ground-breaking work on contractualist theories of justice but it generalizes Rawls’ theory of justice into a more comprehensive account of moral wrongness. The basic idea behind Scanlon's form of contractualism is the Rawlsian idea that moral principles are principles that everyone can agree are justified. Scanlon develops this idea so that the focus is not on universal agreement but rather on the idea of ‘principles that no one could reasonably reject given the aim of finding principles that others, similarly motivated, could also accept’ (Scanlon, What, p. 33). With more of its nuances, here is a fuller statement of his main idea: ‘An act is wrong if its performance under the circumstances would be disallowed by any set of principles for the general regulation of behavior that no one could reasonably reject as a basis for informed, unforced agreement’ (Scanlon, What, p. 153). The key concept here is of course the ‘reasonable’. All of the real normative work will depend on what are and what are not grounds for reasonable rejection of a set of principles.

For all that Scanlon has to say, I just do not see the argument that it would be reasonable to reject the two principles of Kantian consequentialism. To review, the first principle is a requirement to promote the flourishing of rational agency. The second principle is a requirement to promote the best possible circumstances for agents to effectively advance their ends. The first principle constrains the second and thus reflects the Kantian priority of rational nature over happiness. These principles reflect the special value of rational nature as the source of the value of one's ends and the equal status of all persons as sources of valid claims. They do not presuppose that welfare (or well-being) is the only substantive value and thus do not involve a simple aggregative conception of the human good. Kantian consequentialism thus seems to provide a reasonable basis for unforced agreement among persons.

Scanlon himself defends an account of substantive values that is supposed to provide a basis for determining what is and what is not reasonable (Scanlon, What, p. 194). Simon Blackburn has pointed out that this aspect of Scanlon's theory is puzzling and problematic. Briefly, if substantive values provide the grounds for determining what counts as a reasonable rejection of a principle, then the contractualism does not seem to do any normative work. This problem with Scanlon's theory is known as the ‘redundancy objection’.Footnote ²⁰ Moreover, if Scanlon's view is indeed supposed to be a form of Kantian rationalism, then it faces the additional problem of the status of the substantive values themselves.

Michael Ridge has argued that the redundancy objection to Scanlon's contractualism misses the mark in an important and fundamental way.Footnote ²¹ The objection assumes, Ridge argues, that the grounds for reasonable rejection of a principle are agent-neutral but on Scanlon's account the grounds for reasonable rejection are, and must be, agent-relative.Footnote ²² If the grounds were agent-neutral, then the critics would be right that the contractualist element is redundant and adds nothing. But since the grounds are instead agent-relative, the contractualist element provides a model procedure that brings together, as inputs, the personal, agent-relative reasons of distinct persons, and generates impartial reasons that we can share, and thus moral principles that establish ‘what we owe each other’.

Whatever the overall merits of this proposal,Footnote ²³ it simply does not help with the project of providing a Kantian basis for a non-consequentialist normative theory that includes deontological constraints. Let us look at it more closely. Ridge, following Nagel, explains that there are three general types of agent-relative reasons: ‘reasons of autonomy’, which are reasons that are based in an agent's particular projects and central goals; ‘reasons of special obligation’, which are moral reasons based on one's special relationships to family and friends or particular commitments that one has made to others; and ‘deontological reasons’, which are constraints on the actions that one can take to promote one's projects or any agent-neutral good. The rough idea is that these agent-relative reasons provide the substantive grounds for the reasonable rejection of principles. Let's say one person, agent A, proposes a principle that advances A's goals and another person, agent B, can reasonably reject the principle based on B's agent-relative reason. B's reason for rejecting the principle is not the same ground as A's reason for accepting this rejection, because B's reason is agent-relative. A's reason must instead be that A recognizes that B's agent-relative rejection is reasonable, all things considered, and it is this recognition of reasonableness that is the moral reason to refrain from acting on the principle in question. The contractualism is thus not empty and redundant for it is the procedure that combines agent-relative reasons and shows, using the (substantive) notion of reasonable rejection, which actions are permissible and which actions are prohibited – so the argument goes.

The obvious problem with this solution to the redundancy problem is that it must assume a broad class of moral reasons and obligations that are prior to the contractualist procedure and thus lack a basis of their own. The promise of contractualism was that it was supposed to provide a Kantian justification of deontological constraints, and perhaps also options (or a prerogative) to give disproportionate weight to our own projects and relationships. Ridge solves the redundancy problem by assuming as his starting point that there are agent-relative options, special obligations and constraints. Even if Scanlon's contractualism is not redundant and empty, on Ridge's interpretation it does not do the hard work that it was supposed to do. Perhaps, however, there is a more promising approach. Perhaps Ridge helps himself to too much in incorporating all three types of agent-relative reasons.

Scanlon's contractualism should set aside at the start agent-relative moral reasons, that is, special obligations and deontological reasons, and start only with the non-moral agent-relative reasons of autonomy. On this revised interpretation, the grounds for reasonable rejection can appeal only to the fundamental projects and commitments of the contractors. The problem for the deontologist, however, remains. We still need an explanation of how this procedure gives rise to agent-relative ‘reasons of special obligation’ and ‘deontological reasons’. Such a claim is utterly ad hoc, unmotivated and unjustified. Since the only agent-relative reasons that are inputs to the contractualist procedure are ‘reasons of autonomy’ – individual commitments to projects, relationships and other personal goals – the moral reasons must piggyback on these types of reasons alone. One obvious proposal is that reasonable grounds for the rejection of principles must be based on the assumption that there is a corresponding agent-neutral reason for every agent-relative reason. The result would be a principle of reasonable rejection that gives impartial equal consideration or weight to each agent's ends. So interpreted, it is hard to see why this does not result in some form of consequentialist principle. What we have here is a familiar argument for utilitarianism – with the standard concerns over whether the fundamental principle will be a maximizing or satisficing principle, and of course whether it will be simply aggregative or also include distribution-sensitive considerations based on the priority of the worst-off or egalitarian values.Footnote ²⁴ In short, contractualism so interpreted seems to generate some form of aggregative and/or distribution-sensitive consequentialism. I believe that the closest Scanlon comes to addressing this alternative is in his discussion of ‘welfare contractualism’. His rejection of welfare contractualism, however, brings us back to the problem with which we started.

In his argument against ‘welfare contractualism’, in which only considerations of well-being provide reasons for reasonable rejections, Scanlon appeals to substantive moral judgments as an additional basis for the ‘reasonable rejection’ of proposed principles. As should now be clear, the problem for Scanlon's account of morality is that substantive moral judgments about whether a set of principles can reasonably be rejected will reflect the individual's own prior moral views about which acts or principles are right or wrong. The judgment that an action is wrong cannot be the agent's reason for rejecting the set of principles in question, for then the contractualist procedure does not establish the wrongness of the action. It seems clear, however, that our substantive judgments of the reasonableness of a set of principles do indeed reflect our views of what is or what is not permissible. If Scanlon's procedure is to provide an independent account of moral wrongness, we need the contractors to set aside their prior views – their ingrained intuitions – about rightness and wrongness. Scanlon's idea of reasonable rejection, however, really invites us to appeal to our prior moral intuitions about what it is reasonable to reject. The resulting contractualist model of moral wrongness thus does not provide an independent justification of common moral intuitions. An action is wrong if it can be reasonably rejected and the basis for reasonable rejection is a prior intuition about what is wrong. The contractualist element does not seem to do any real work other than obscuring and hiding the appeal to our moral intuitions. It would thus be philosophically preferable to just directly appeal to these deontological intuitions.

Scanlon does discuss and reject the utilitarian idea that reasonable rejection of principles must appeal to considerations of well-being alone.Footnote ²⁵ By focusing on utilitarianism, however, Scanlon's argument here does not address adequately more complex forms of pluralistic consequentialism, or distribution-sensitive consequentialist theories, or a consequentialist theory with a two-tiered theory of value that gives priority to development and flourishing of rational nature over considerations of well-being alone.

The claim that consequentialism in general can be reasonably rejected may nonetheless seem a straightforward matter. Surely, it will be argued, I can reasonably reject a principle that requires my innocent self to be sacrificed for the sake of the greater good of others. However, the point is that we find the rejection of this principle reasonable only because we already accept the moral constraint on harming innocents. If the inputs are simply the agent-relative reasons of all persons affected, then we have the competing claims of (1) the persons that would be harmed and of (2) the persons needing aid. We thus need an explanation of why the one type of claim gives rise to reasonable rejection and the other does not, and so far we have been given no such explanation.Footnote ²⁶ In the next section, we will consider Kamm's account of the inviolability of persons as a basis for filling this gap.

Scanlon also tries to defend his use of substantive judgments by arguing that moral principles are not simple rules and thus we must rely on ‘significant appeals to judgment’ (Scanlon, What, p. 246). In response to this obvious point, we must distinguish the clearly necessary use of judgment in applying principles from a reliance on substantive moral intuitions for which one can provide no defense. The former is obviously necessary. But to insist that substantive moral judgments of what is right and wrong are necessary abandons the promise of contractualism as a basis and justification of moral judgments. And it does so simply because one cannot otherwise account for one's intuitions within the limits set by one's own account. No matter how strongly such ‘judgments’ are felt, others can surely reasonably reject them.Footnote ²⁷

Many Kantians insist that when we are reflecting on what we can take to be a reason, we must also recognize the value of rational nature, or rational reflection, itself, as the source of all reason-giving values, and thus we must treat rational or reflective nature as an end-in-itself. Scanlon seems to implicitly accept the Kantian model of autonomy of the will as the source of all moral authority. After all, his whole contractualist theory is based on an ideal of justifiability to others and on the importance of uncoerced agreement. At any rate, despite Scanlon's own more cognitivist pretensions, his arguments seem to support only a proceduralist model of reflective endorsement. It follows from this type of model that an agent's reasons for rejecting principles must be based on the basic ends that, upon reflection, they set themselves. As many have pointed out, it is not clear how this alone will lead to a point of agreement that will ground shared principles. Kantian contractualism, however, can include the value of rational nature as an end-in-itself, and thus as an internal constraint on the setting of ends. One possibility, then, is that the shared commitment to the special value and dignity of persons can provide a basis for balancing individual ends, or agent-relative reasons of autonomy, and thereby provide a shared conception of reasonable rejection. Specifically, it is reasonable to reject principles that do not respect the special dignity of persons. Since Kantian consequentialism, with its priority of rational nature, offers one (reasonable) interpretation of respect for the dignity of persons, it is reasonable to conclude that these are ‘principles that no one could reasonably reject given the aim of finding principles that others, similarly motivated, could also accept’ (Scanlon, What, p. 33).

I have argued first that Scanlon's contractualism, as a result of its reliance on substantive values and intuitions, does not succeed as a distinctly Kantian deontological alternative to Kantian consequentialism. I have also suggested that Scanlon provides no rationale, except a direct appeal to deontological intuitions, for rejecting consequentialist principles. He thus simply does not provide a Kantian contractualist alternative to consequentialism. Having considered and rejected the indirect contractualist approach, we will now consider more substantive conceptions of respect for the dignity of persons.

DIGNITY, STATUS AND THE INVIOLABILITY OF PERSONS

The substantive approach is less straightforward than the contractualist approach because it can still recognize the value of consent. Substantive theories, however, also use the concept of dignity as a substantive value or status that functions independently from consent and that limits the legitimating power of consent. To bring out the difference in these two approaches consider the familiar tool of an initial fair choice situation, like Rawls's Original Position. Leaving aside questions about the overall construction of the free choice situation, the justification for the veil of ignorance, or the motivational postulates, a pure proceduralist conception will treat the outcome of the procedure as justified because it is an agreement between free and equal persons and thus respects the dignity of each. For a pure proceduralist, the contractors do not aim to respect dignity as a separate concern; we respect dignity when we act in accordance with and are motivated by the principles that are the outcome of the procedure. Respect for dignity is fully captured by the concern to regulate one's conduct by the principles that result from the procedure itself. If, however, the contractors are motivated by a concern to respect the dignity of humanity, and this concern places substantive limits on what they can reasonably consent to, then dignity also functions as a substantive value or principle that limits the legitimating force of the procedure itself.Footnote ²⁸

As we have seen in discussing Scanlon's theory, however, appeal to substantive values and principles raises a problem: if the additional substantive value is not itself either reflected in the procedure or itself constructed, what, then, is it its source? It would seem that any substantive value, other than the direct value of protecting and promoting the capacity for autonomy itself, would require a commitment to a form of substantive moral realism that is incompatible with Kantian rationalism and, indeed, with the central Kantian idea of autonomy as the source of morality and of the dignity of humanity. For now we are operating with a conception of human dignity that transcends the autonomy of the will as captured by principles internal to deliberative rationality. Paradoxically, if dignity functions as an independent substantive value, we would have a form of heteronomy that undermines the autonomy of the will and thus the ‘higher vocation’ and dignity of humanity.

Recall that Kant uses the concept of dignity to mark off the special status of rational nature. We have seen that the value that we must place on our rational nature is indeed a higher value in the sense that it is conceptually prior to all other values. It is also a necessary or objective value in that its value is not contingent on the individual subjective judgment of a particular person (and in this sense it has no price in the currency of inclinations alone). According to the Kantian conception of value, something is valuable because it is the object of free and rational choice. Our choices do not track objective value, as metaphysical realists maintain; rather, something is valuable because it is the object of rational choice. So, as the source of all contingent value, rational nature has a special value and it is this value that is the basis of human dignity.

It is indeed standard philosophical practice to use Kant's conception of human dignity to argue against consequentialist conceptions of ethics and for a deontological conception of the inviolability of persons (roughly speaking, inviolability is the principle that it is wrong to violate persons by harming them). One especially noteworthy version of this approach appeals to the association between having the status of inviolability and having a special dignity. Frances Kamm, in particular, has argued that by treating persons as inviolable we recognize the special status of persons. The idea here is that the recognition of an inviolable status reflects the higher value of persons and the special dignity of humanity.Footnote ²⁹

So the question is whether Kantian consequentialism fails to express a due regard and recognition of the dignity of persons because it does not adequately recognize the inviolability of persons. In response, we must first note that the priority of the value of rational nature entails that a person's rational nature cannot be violated to promote the happiness or satisfaction of other persons. Let us call the prohibition on violating the fundamental rights, liberties and security of a person in order to advance the contingent, subjective interests of others, first-order inviolability. This is the Kantian consequentialist conception of inviolability.

We must ask as well whether it is also wrong to infringe rights in order to prevent the rights of others from being more extensively violated. (An infringement of a right is an act that would be a violation if it were not justified by special circumstances.) Let us call the prohibition on violating rights to protect the rights of other persons from being violated, second-order inviolability. Kamm argues that second-order inviolability is a higher status than first-order inviolability, and thus it provides a fuller recognition and expression of the special status and dignity of persons. It is not at all clear, however, why second-order inviolability marks a higher status than first-order inviolability. After all, it is a predictable consequence of respect for second-order inviolability that, in a world where some people will violate rights, persons are more likely to have their rights violated as a result of having this additional status respected. Since second-order inviolability disallows actions that would minimize violations of persons (indeed this is its point), it countenances more violations than does the status of first-order inviolability and a principle of minimizing violations.

In one sense Kamm is right that the refusal to violate a person's rights, even when this would prevent other more extensive violations of the rights of other persons, treats that particular person at that moment as more inviolable. Her creative (and indeed brilliant) argument emphasizes that inviolability is a status and one has this status even if one's rights are in fact violated. Even those whose rights are violated still have the status of inviolability. So second-order inviolability, which also prohibits violations that minimize violations, adds an additional status of inviolability. Now, this may cause us to think that this second-order principle makes a person less likely to be violated; for the person will not be violated even when it is necessary to prevent others from being violated. But this narrow perspective neglects the other innocent persons who are nonetheless still violated instead. In fact, since it prohibits minimizing violations of rights, everyone's rights are probably more likely to be violated in a world that includes respect for the status of second-order inviolability. Since the second-order principle countenances a world where more people's rights are violated, it is not at all clear that it expresses a greater overall respect for the dignity of persons.

The problem here is that second-order inviolability permits the rights of others to be violated when this could be prevented. Kasper Lippert-Rasmussen has described this as ignorability because it allows us to intentionally ignore preventable violations of rights.Footnote ³⁰ Ignoring preventable rights violations fails to respect the dignity of persons and so our status is increased if we have the status of unignorability, that is, if others are required to prevent rights violation. It follows that, if there is a requirement to minimize rights violations, then we lose second-order inviolability but we gain the status of unignorability. Since first-order inviolability and unignorability together prohibit rights violations unless infringing a right better secures the rights of all, our rights are more secure overall if rights violations are to be minimized (that is, if we do not have the status of second-order inviolability).

Now, if the particular rights (and liberties) promote the flourishing of rational nature, then the minimization of rights violations also preserves and promotes the flourishing of rational agency. Since the goal, in cases in which the rights of some are infringed to prevent the more extensive violation of the rights of others, is the shared goal of protecting, preserving and honoring the rights of all, then infringing a right in these specific circumstances actually clearly expresses a respect for the rights, inviolability and dignity of persons. From this perspective, the status of second-order inviolability actually conflicts with the special value and dignity of rational nature.

On the other hand, if we stay on the level of mere intuitions about status, then I do not see how we are to resolve the conflict between the claims of increased second-order inviolability and the claims of unignorability.Footnote ³¹ Indeed Kamm admits that the unignorability principle (which she calls ‘saveability’) also seems to express a heightened status. She grants that ‘it does seem that if a creature is significant, it is wrong to abandon it and that some costs should be imposed on others to help it’. This is too modest, however. We have even greater ‘status’ if we believe that rational nature has such special value that we are committed to minimize the likelihood of rights violations, even when one's own rights may be infringed to better secure the rights of all. Intuitively, this involves a striking expression and recognition of the special value and dignity of persons. So, the status of inviolability and unignorability both have intuitive appeal. We have an additional Kantian basis, however, for embracing first-order inviolability in that it directly respects the rights that preserve and promote the flourishing of rational nature. Similarly, all sides agree that rights violations by others should not be ignored and we should prevent others from violating these fundamental rights. The principle of unignorability is thus also directly grounded in the Kantian value of rational nature. The additional status of second-order inviolability has no similar clear Kantian basis and it makes our rights less secure; consequently, there is no clear sense in which it expresses or recognizes the special value of rational nature.

In response to this type of criticism, Kamm has suggested that second-order inviolability is more important than ‘saveability’ because it expresses respect for a ‘rational will, whose consent we must seek when interfering with what a person has independently of imposition on us’.Footnote ³² If this is not simply begging the normative question at issue, then it returns us to the contractualist approach that has already been considered and found wanting. Indeed, we turned to Kamm's dignity/status argument to supplement and fill the gap in the contractualist case against consequentialism. When pressed, contract theorists appeal to Kamm's status argument, yet when Kamm is pressed she abandons appeals to status and appeals to contractualism. The consequentialist faces a deontological shell game of changing arguments, and this simply further reinforces the conclusion that second-order inviolability, and thus the classic deontological constraint, is really a mere intuitive principle forever in search of, but never achieving, a principled Kantian justification.

I do not deny the strength of the intuition that supports a principle of second-order inviolability and that opposes the minimization of rights violations: we recoil at the thought of harming the innocent – even in cases when it seems necessary to prevent similar or more serious harm to others. All who share this intuition immediately think of even second-order violations of persons as simply wrong. But it is this prior and unmediated judgment that these types of violations are wrong that provides the basis for the belief that they are incompatible with respect for the dignity of persons. And, however appealing and deeply felt these judgments may be, Kantian rationalists and consequentialists agree that substantive moral intuitions do not necessarily track moral truths. Substantive moral intuitions are too often simply a reflection of the widely held moral beliefs of one's particular culture, and thus these internalized intuitive beliefs must be vindicated and justified by critical reflection. (Although this is not the topic at hand, I believe that a sophisticated consequentialist can provide an adequate account of common sense morality.)Footnote ³³ More recently, it has been argued that deontological intuitions are based simply on innate feelings, which were more suitable and beneficial in our evolutionary past. On these accounts, ironically, Kantian deontology is founded on contingent evolutionary psychology and not reason at all.Footnote ³⁴

Kantian ethical theories must avoid appeals to substantive realist principles or they betray their rationalist foundations. The concept of the dignity of humanity, however, seems to encourage us to import our substantive moral judgments into our concept of human dignity. What happens, I believe, is that we use our sense of what is right and wrong to determine what is and what is not consistent with respect for dignity. If we treat someone wrongly, we do not respect that person's dignity. So we naturally (but mistakenly) use our substantive intuitions about what is right or wrong to decide what does and what does not respect dignity. The problem is that when we use the concept of dignity in this substantive way, it cannot provide the basis for our judgment that the act is wrong. On the contrary, a prior judgment that the act is wrong has become our basis for the judgment that the act does not respect human dignity.

The argument for Kantian consequentialism must also avoid substantive values and principles that are not derived from the internal perspective of rational agency. Kagan's ‘Consequentialism for Kantians’ is based on the purely formal considerations that are supposed to support Kant's universal law formulation of the categorical imperative. My own argument emphasizes the special value of rational nature. Some may argue that the consequentialist, like the deontologist, must rely on additional substantive values and principles. It is a hollow victory for the Kantian deontologist, however, if we must conclude that neither consequentialism nor deontology follows from Kantian rationalism. If this is the case, it also marks the failure of Kantian ethics as a distinctive method capable of generating substantive moral conclusions. Indeed, if Kantian rationalism is consistent with consequentialism and deontology, then the choice of substantive moral principles simply reflects a prior intuitive receptivity to deontology or consequentialism. If this is so, the deontologists should abandon the baseless claim that consequentialism fails to treat persons as ends, and thus does not respect the dignity of persons. We have instead nothing more than two competing interpretations of the idea of treating persons as ends and of respect for human dignity. In short, if all that can be said against the Kantian consequentialism is that it is built on an initial consequentialist bias, the corollary holds that the deontologists, despite their noble efforts, have provided no rationale whatsoever, beyond the simple bias of their intuitions, for deontological constraints.

In fact, however, we are not faced with a simple stalemate and standoff of mere competing intuitions. The deontologist is actually in the weaker position and has the burden of proof. If Kantian consequentialism is based on principles shared by the deontologist and the consequentialist, as I have argued, and the deontologist is defending an additional substantive principle, then the consequentialist may fairly ask: what is the distinctively Kantian basis for this additional principle?