1 Although he does not acknowledge it, Kant entertains different conceptions of the highest good. According to the traditional view held by commentators, Kant thinks of the highest good as happiness proportioned to virtue.Footnote 1 This idea, which I refer to as the proportionality thesis, is given extended treatment in the Critique of Practical Reason (CPrR 5: 107–32). However, John Rawls (Reference Rawls2000) and Andrews Reath (Reference Reath1988) point to places where Kant thinks of the highest good as the maximization of happiness and virtue (CJ 5: 453; TP 8: 279). They argue that this version of the idea – which I refer to as the maximization thesis – is Kant's mature view. In this paper I will argue against Rawls and Reath in favour of the traditional view. The positions adopted by Rawls and Reath do not significantly differ, but Reath develops the argument at greater length; my refutation is, therefore, directed against him. As it is difficult to develop a consistent picture of the highest good from Kant's many discussions of the topic, I have opted to construct a view of the idea that is internally consistent, and which generally fits with Kant's thoughts on the synthesis of happiness and virtue.

The discussion proceeds as follows. In section 2, I do some unpacking of the two ideas, and show how they can generate opposing evaluations of certain situations. This makes it necessary to determine which ought to be considered the highest good. I argue that we cannot do this based on intuitions, and therefore that we need to derive the content of the idea from Kant's theory of practical reason. Section 3 focuses on what I refer to as the ‘futility argument’, which is Reath's argument that proportionality cannot be the highest good because it cannot be realized by purely human means. I argue, first, that this is based on a problematic reading of Kant's discussion of objects of the will in the second Critique; and second, that Kant's theory of ideas (Ideen) implies that all ideas are unrealizable by purely human means. In section 4, I criticize Reath's derivation of the maximization thesis; then, in section 5, I show how the proportionality thesis responds to a dualism between prudence and morality. In section 6, I criticize an argument by Jeffrie Murphy that is similar to Reath's futility argument. Murphy argues that there is no course of action open to us to promote proportionality. This discussion provides an opportunity to consider how agents might go about promoting the realization of the highest good. In section 7, I discuss the highest good's place in practical reason. Finally, in section 8, I discuss how the two theses can be made compatible.

2 Kant often portrays the highest good as a standard of evaluation for ‘impartial reason’ (G 4: 393; CPrR 5: 110; LE 27: 24). This implies that we should be able to use the idea to evaluate situations; we should also be able to use the idea to rank as better or worse different situations, or changes in situations. We will see however, that the two versions of the idea can generate opposing evaluations.

The proportionality thesis holds that people ought to be happy to the degree that they are virtuous, unhappy to the degree that they are vicious. This generalizes the intuition that there is something wrong when virtuous people, such as Job, suffer. As such, it is a formal articulation of the idea of ‘just deserts’. If we adopt the proportionality thesis as a standard of evaluation, then we will positively evaluate any situation in which there is proportionality between happiness and virtue. Now this seems to imply that we should positively evaluate any change that establishes, or more closely approximates proportionality; but this could include situations in which virtuous but miserable people, such as Job, become more vicious. As this implies a violation of the categorical imperative, Kant can only maintain the proportionality thesis if it does not carry this implication. When we consider, in sections 4–8, how the proportionality thesis arises in the wider context of Kant's practical philosophy, we will see how this is, in fact, not implied by the proportionality thesis; for now I will pre-empt that discussion and say that the proportionality thesis only implies a positive evaluation of situations in which everyone gets their just deserts.

The instantiation of the proportionality thesis would, if necessary, increase the happiness enjoyed by the virtuous until it is proportional to their virtue; but it would also decrease the happiness of vicious people to the point at which it is proportional to their vice. If decreasing someone's happiness is understood as a form of harm, then the proportionality thesis implies that it is morally proper that the vicious are harmed in proportion to their viciousness.Footnote 2 In contrast, any reduction in the level of happiness could not be positively evaluated from the perspective of maximization; evaluations determined by the maximization thesis will be positive only if there is some total increase in the level of either, or both, values. The maximization thesis can only be instantiated in a perfect world of complete happiness and virtue, whereas the proportionality thesis can be instantiated in less perfect worlds.

The perfect world of maximal happiness and virtue also instantiates perfect proportionality, but this idea does not tell us how to evaluate the relative merits or demerits of less than perfect worlds. For example, a world in which everyone was completely vicious and devoid of happiness – call it a universal hell – would instantiate the proportionality thesis. But this is also the worst possible world from a maximization perspective. An alternative and completely unjust world, in which the virtuous are miserable and the vicious happy, would be better than a universal hell from a maximization perspective; but it would be the worst possible world from a proportionality perspective.

Our intuitions about these cases do not help us to adjudicate between the two versions of the idea. It seems counterintuitive that a universal hell instantiates the highest good, but this presupposes that the highest good must include some positive amount of happiness, regardless of virtue. But since this is not necessitated by the idea of proportionality between two values, it cannot be presupposed without begging the question in favour of a value maximizing conception of the highest good, and against the proportionality thesis. It also seems counterintuitive to take moral satisfaction in moving from a universal hell to a completely unjust world. But a completely unjust world contains more happiness and virtue than a universal hell, and therefore it is closer to the ideal of maximization. Since the central commitment of the maximization thesis is just to increase both values to the maximum, it provides no basis from which to argue that an increase in happiness should be restrained in order to maintain proportionality with virtue. That the maximization thesis is not committed to the general redress of unjust suffering is admitted by Reath. He argues that it is consistent with the highest good that some virtuous people might have to suffer for the sake of realizing the maximization thesis (Reath Reference Reath1988: 603). Reath's admission that this ‘might seem unfair’ invalidates the maximization thesis's claim to be the highest good from the proportionality perspective; however, we cannot make too much of intuitions about just deserts without begging the question against the maximization thesis and in favour of proportionality.

Considering scenarios like these, Kant notes that neither amounts to the idea of a perfect world, the ‘highest created good’ (LE 27: 248). But his failure to distinguish between the proportionality and maximization theses leaves open the question whether the primary focus of evaluation should be on the proportion between, or the total amount of, the two values. But Kant does not simply adopt an idea of the highest good; rather, he derives it from his theory of practical reason, and we have not yet considered this context. The path this derivation takes is not clear, however, and Reath's reconstruction of it leads him to argue against the proportionality thesis and in favour of maximization. Let us then consider his argument.

3 The textual locus for Reath's argument is the second chapter of the Critique of Practical Reason, in which Kant discusses ‘objects of the will’. An ‘object of the will’ is the representation of an action's goal or purpose, an end of practical reason. Broadly speaking, practical reason characterizes objects of the will as ‘goods’; thus Kant's discussion of objects of the will is also a discussion of the concept of the good (CPrR 5: 57–8). Kant goes on to formally designate the highest good as ‘the unconditioned totality of the object of pure practical reason’ (CPrR 5: 108).

There is a difference between thinking of the highest good as an object of agency and a principle of evaluation.Footnote 3 When we make something a goal then we must think of the means to achieve it, and whether, or to what extent, it can be achieved. At the centre of Reath's argument is Kant's claim that an object of the will is ‘the representation of an object as an effect possible through freedom’ (CPrR 5: 57). Reath interprets this to mean that a representation can be an object of the will only if ‘one can will the action that would bring it about’ (Reath Reference Reath1988: 596). This implies that ‘only states of affairs that we can imagine as the possible results of human action’ can be objects of the will (ibid. 597). Reath allows that some objects of the will cannot be fully realized; nevertheless, ‘some relationship between the good and human agency is implied when the good is defined as the possible object of a person's moral intention’ (ibid.).

It is noteworthy that Reath thinks of objects of the will as both actions and the ends of action without distinguishing the two. Later in this section, and in section 6, we shall see that this distinction turns out to be significant; but putting that concern aside, the general thrust of Reath's argument is as follows. If we cannot promote proportionality, then it cannot be an object of the will; therefore, it cannot be the highest good. To promote the realization of the proportionality thesis we need to know an agent's degree of virtue, but we cannot know this unless we have noumenal insight. Therefore, we cannot make proportionality the highest good since there is nothing we can do to promote this goal. As the sense of this argument is that the promotion of proportionality is futile from a purely human standpoint, I refer to this as the ‘futility argument’.

Reath goes on to argue that the futility of pursuing an agenda set by the proportionality thesis invites practical reasoners to posit the existence of God, for only a supreme being could bridge the gap between human abilities and the instantiation of the highest good. But Reath thinks that the introduction of God is incompatible with autonomy, so the proportionality thesis is either futile or heteronomous.

Reath derives the futility argument from the following passage in Kant:

This passage has the strange appearance of inconsistency. At the beginning Kant seems to be contrasting the two kinds of determining grounds based on whether the ‘physical possibility’ of realizing the object is a necessary condition for action. In the first case, when an object is the determining ground of the will, the ‘physical possibility’ of realizing the object is a necessary condition for action. But when the moral law, and not an object, is the determining ground of the will, the physical possibility of realizing the object of the will is not a necessary condition for action. However, Kant then seems to retract this by saying that ‘the question is only whether we could will an action which is directed to the existence of an object if the object were within our power’ (italics added). The final clause seems to reintroduce the necessary condition that only if an object is within our power is it rational to will actions aimed at its realization, even if the moral law is the determining ground of the will. But if that is the case then there is no contrast between the two determining grounds: objects that are physically impossible to realize cannot be objects of the will regardless of what determines the will. Hence Reath interprets Kant to be committed to the position that futile ends – ends the realization of which are beyond our physical powers – cannot be objects of the will.

Yet this is inconsistent, not only with the contrast we expect at the beginning of the passage, but also with other important claims made by Kant. For example, the claim that autonomy is when an agent's maxims are ‘principles that contain the determining ground of the will not by their matter but only by their form’ (CPrR 5: 27) implies that we have the ability to determine the will independently of how it is related to an object. Again, in a well-known passage in the Groundwork, Kant admires the good will's ability to pursue morally principled but ultimately futile courses of action:

But Reath insists that ‘it would make no sense to talk about willing an action directed at an object, even hypothetically, unless it were something we could imagine as a result of human agency’ (Reath Reference Reath1988: 597); and there remains the textual evidence that Kant seems to be saying something to justify Reath's interpretation, albeit in a seemingly inconsistent manner.

The inconsistency between this passage and what seems to be Kant's general position on good but futile intentions raises the question whether there is a problem in the translation. The quotation above is taken from the Cambridge edition; Reath cites the Beck translation where the key statement is rendered, ‘the question is only whether we may will an action that is directed to the existence of an object if [making] this [object actual] were under our control’ (CPrR, trans. Beck, 78). We can see that the antecedent upon which Reath bases his position – that of making the object actual – is an interpolation.Footnote 4 But what Beck renders as the ability to make an object actual H. W. Cassirer translates as ‘the power to perform an action’:

Here the expected contrast is realized in the modal considerations relevant to the different determining grounds: When an object determines the will, the modal consideration is the physical possibility or impossibility of realizing the goal. It would be practically irrational to adopt an impossible end when the entire rationale for an act is to realize that end; and in such cases, the futility of any effort is a compelling reason for inaction. But when the moral law is the determining ground of the will, the modal consideration shifts from physical to moral possibility, and the question becomes whether an action is forbidden, permitted or necessary – not whether the end is realizable by means of that action. The sense is that categorically mandated actions should be performed even if they are futile. For example, if it is morally necessary for me to comfort someone grieving over an irreplaceable loss, then I should do so even if I do not believe that I can give any comfort; or if it is morally necessary that I end child slavery then I should take on this project even if there is no prospect that I will succeed in doing so. Similarly, if the moral law commands that I strive to instantiate the proportionality thesis then I ought to do so, even though I know that I cannot achieve this goal.

Kant's claim now looks to be that it does not matter whether a goal prescribed by pure practical reason is physically impossible, for moral agents should perform categorically mandated actions even when they aim at impossible goals. The greater internal consistency of the Cassirer version, and its consistency with Kant's wider position recommend it over the alternative. Since we cannot conclude that Kant's discussion of objects of the will rules out futile goals, the fact that the proportionality thesis cannot be realized by purely human means does not necessarily rule it out of contention for the position of highest good.

A further doubt about Reath's futility argument arises when we consider Kant's theory of ideas. Ideas (Ideen) are representations of ‘the unconditioned’ posited by reason to avoid an infinite regress (A327/B383–4). By definition, the unconditioned cannot be the object of a possible experience because the forms of experience condition all possible experiences; indeed, Kant insists that the word ‘idea’ should be reserved solely for the use of representations that go ‘beyond the possibility of experience’ (A320/B377). Therefore, a representation cannot both be an idea and something that can be empirically instantiated. Kant does distinguish between theoretical and practical ideas, allowing that practical ideas can be ‘given in concreto, though only in part’ (A328/B384); that is, practical ideas can be used as goals for which we strive. Nevertheless, all attempts at realizing them will be ‘defective’ (A328/B384).

Now, we might think that it is open to Reath to argue that the maximization thesis can be realized, either because it is not an idea, or because it is a special kind of idea. This brings us to consider how Reath envisions its realization. He begins his account by referring to the brief discussion of a ‘self-rewarding morality’, in which Kant writes that ‘freedom, partly inspired and partly restricted by moral laws, would itself be the cause of general happiness’ (A809–10/B837–8; Reath Reference Reath1988: 615). Reath then suggests that virtuous individuals, acting from self-interest in the context of the right kind of society, would ‘take a positive interest in the happiness of others, as well as a concern for the common good’. ‘It is not implausible’, Reath continues, ‘that individuals would by and large be successful in achieving their ends in a system of this sort.’ That is, ‘an effective system of conduct guided by the Moral Law would lead to the happiness of its members’ (Reath Reference Reath1988: 616). Kant further develops this idea, according to Reath, in his idea of the ‘ethical commonwealth’ in the Religion.

There are, however, problems with this account. First, Kant denies that the relevant ideas can be realized phenomenally, or through human efforts alone. In the first Critique, he writes that the connection between morality and happiness ‘cannot be cognized through reason if it is merely grounded in nature’, and that a self-rewarding morality is ‘only an idea’ that ‘may be hoped for’ (A809–10/B837–8; see also CPrR 5: 113). In reference to the components of the highest good, Kant thinks of virtue as an idea of reason, and that ‘[h]uman virtue is always imperfect’ (LR 28: 994; see also G 4: 407–8). Happiness is an idea of the imagination, rather than of reason, which means that it cannot be clearly conceived, and therefore it is difficult for agents to effectively direct their activity towards its realization (G 4: 399). Thus Kant writes that ‘[w]e cannot direct our actions according to an idea of happiness’ (LR 28: 1057–8; italics in original). This does not mean that we cannot pursue happiness, but Kant argues that the rational pursuit of happiness will always be counterproductive (G 4: 395). Finally, Kant explicitly states that the realization of an ethical community requires divine assistance:

The central problem, however, is that Reath's idea requires that there be a causal connection between the universal exercise of virtue and the realization of the maximal happiness. There are two reasons for this. First, since the highest good is an object of practical reason, it must be ‘a state of affairs that we can envision as the result of our conduct’ (Reath Reference Reath1988: 607). Second, there cannot be a ‘naturalistic’ account of the realization of the maximization thesis unless there is a causal connection between virtue and happiness. Reath's emphasis on the creation of the right kind of society does not alter this requirement; it merely adds some detail to the connection. To consider the possibility that universal virtue and good social systems might secure general happiness is to entertain the idea that virtue could be a sufficient condition for happiness. But there is no phenomenal basis for the belief that such a connection actually exists. The tentativeness of Reath's language in this context is noteworthy; the idea, he writes, ‘does not appear to rest on unreasonable assumptions’, ‘it is not implausible’, and so on. Reath's discussion is a matter of pure speculation. We might seek to bolster this speculation with reference to Kant's essay on the Idea for a Universal History with Cosmopolitan Intent, but we have to note that, once again, Kant regards the thought that there is a teleology in history leading to ‘a perfect civil union’ as an ‘idea’. He insists that it stands only as a ‘standpoint’ – which we might gloss as an interpretation – that cannot displace ‘true empirical history’ (I 8: 31). Kant further argues in the Critique of Judgment that teleological interpretations of phenomena are ‘reflective’ rather than ‘constitutive’ judgments (CJ 5: 379–81, §67). The connection between virtue and happiness that is a necessary part of Reath's conception of the highest good is, therefore, an idea. This rules out the possibility that there can be a ‘naturalistic’ account of the realization of the maximization thesis, and makes it also subject to the futility argument.

4

In addition to his argument against proportionality, Reath gives a derivation of the maximization thesis from Kant's normative theory. On the traditional view, the proportionality thesis is meant to overcome a value dualism between the ‘natural good’ of happiness (promoting well being and avoiding woe), and the moral good (good as opposed to evil) (CPrR 5: 59–60). Reath argues, however, that there is no such value dualism.

As we have seen, Reath thinks that only objects that can be realized through purely human means are objects of practical reason. An object of pure practical reason then, is any end ‘that could result from the moral use of freedom’ (Reath Reference Reath1988: 596). Since objects of the will are ‘goods’, objects of the moral use of freedom – objects of pure practical reason – are moral goods. But ‘the moral use of freedom’ is ambiguous, it can be construed either as actions based on obligatory or permissible maxims. If it refers only to the determination of the will by obligatory maxims, then the morally conscientious pursuit of happiness is a non-moral good. But if the moral use of freedom includes all permissible actions, then it includes the permissible pursuit of happiness. Reath adopts the latter interpretation, namely, that the morally conscientious pursuit of happiness is an object of pure practical reason.

According to Reath, the highest good is the idea of the ends of a morally conscientious will ordered into a systematic whole. The basic and formal conception of the highest good is just the ideal ordering, and we get content by injecting human interests into the formal structure. Since this content is the pursuit of happiness, the highest good is the structuring of permissible ends pertaining to the pursuit of happiness into an ordered whole. Ends are vetted for their moral permissibility upstream of this ordering by the moral law, and we are left with that part of the pursuit of happiness that is consistent with a moral use of freedom. Since permissible happiness is now understood as a moral good, there is no value dualism between the pursuit of happiness and moral obligation; there is, therefore, no need to bring about a synthesis of happiness and virtue. There is a duality of values, but not a dualism. Since the maximization thesis is just an extension of Kant's notion of the object of pure practical reason, and the notion of an object of pure practical reason already contains within it the proper relationship between virtue and happiness, the traditional justification for the proportionality thesis, which is to establish this relationship, falls away.

Now Kant claims that ‘the good’ is the ‘necessary object of the faculty of desire’, and ‘evil’ is the ‘necessary object of the faculty of detestation’ (CPrR 5: 58). As it is not the case that people are averse to committing evil actions, and there is nothing to stop someone perversely willing an evil act precisely because it is evil, the sense of ‘necessary’ here must be deontic. The good is what we ought to want, and we ought to want to avoid evil. Therefore, in this context, ‘desire’ should be taken to refer to the ‘higher faculty of desire’, namely, rational desire (CPrR 5: 24–5). Now according to Reath, the good contains the permissible pursuit of happiness; therefore, the permissible pursuit of happiness is a necessary object of the faculty of desire and an obligatory end. But Kant argues that it is ‘self-contradictory’ to assert that ‘[we are] under obligation to promote [our] own happiness’ because we cannot have an obligation to do something that we naturally want to do (MM 6: 386; see also CPrR 5: 93).Footnote 5 Reath cannot avoid this conclusion by separating the permissibility from the pursuit of happiness because it is the permissible pursuit of happiness, and not anything else, that is, according to him, the object of pure practical reason. Thus Reath is left with no distinction between permissible and obligatory ends.

A further problem with Reath's argumentative strategy is that it can be used to generate a version of the proportionality thesis. As previously mentioned, Reath tries to deduce the maximal conception of the highest good from Kant's discussion of the objects of practical reason. In this context, Kant characterizes good and evil as the sole objects of practical reason, and as necessary objects of rational desire and detestation, respectively. This designation suggests that an impartial rational spectator desires that agents will be successful at achieving their ends, insofar as they are virtuous; and hopes that they will fail insofar as they are vicious.

Now, the determining grounds of the legislative will (Wille) are not chosen, which means that good and evil are always options for finite rational beings. From a phenomenal perspective, the choice to act on the basis of a practical inference shows that an agent subscribes to certain practical principles. A person is not virtuous or vicious from a phenomenal perspective because she has adopted or rejected the moral law; rather, she is virtuous or vicious insofar as her choices demonstrate a commitment to the moral law. As we are constantly making choices, the will can become more or less virtuous or vicious, and the moral quality of an agent unfolds over the course of a lifetime. Since agents may choose rightly or wrongly at any stage, moral evaluation must track this development. The moral evaluation of an agent will, therefore, fluctuate in relation to the choices made by that agent. Thus we can think of moral evaluation in terms of a proportional rational attraction towards, or aversion from, the moral quality of the choices made by agents. Kant defines happiness as everything going as one wills and wishes (CPrR 5: 124); therefore, proportionality between happiness and virtue is instantiated if an agent fails in the evil, and is successful in the good, that she wishes and wills. The impartial rational spectator, who wants virtue to succeed and vice to fail, will pronounce situations in which an agent is only as happy as she is virtuous as morally optimal. Therefore we can follow Reath's lead and arrive at the conclusion that the highest good for an agent is when she is successful in fulfilling her good intentions, and fails at fulfilling her bad intentions, which is just happiness in proportion to virtue. This version of the proportionality thesis contains a notion of just deserts – good intentions deserve to succeed, evil intentions deserve to fail. It also disallows the positive evaluation of the possibility, raised in section 2, that proportionality could be achieved by increasing vice when there is a dearth of happiness.

5 As mentioned at the outset, the traditional reading is that Kant developed the idea of the highest good to overcome the ‘dualism’ that emerges between virtue and happiness as a consequence of the derivation of the categorical imperative. Kant's comment that virtue and happiness are ‘heterogeneous’ (CPrR 5: 112) implies that they are, in some sense, opposed to each other. Yet there is a categorical obligation to strive to be virtuous, and reason ‘has a commission from the side of … sensibility’ to pursue happiness (CPrR 5: 61). Thus neither can be eliminated from the concerns of practical reason. Kant, the traditional story goes, needs a mechanism to bring them together, which he finds in the proportionality thesis. However, Reath argues that we only need the moral law to establish the proper relationship between virtue and happiness, rendering the proportionality thesis otiose. The question is, then, whether virtue and happiness are opposed in such a way as to create a value dualism that requires a synthesizing idea. In this section, using the prisoner's dilemma, I will argue that it is rational to think that the virtuous person is less likely to achieve happiness; and that this creates a value dualism which justifies the synthesis of virtue and happiness articulated in the proportionality thesis.

Happiness and virtue are different kinds of determining grounds of the will. Happiness is an empirical determining ground and virtue is a pure determining ground. But a determining ground of the will is pure if and only if it contains no empirical element; therefore, virtue and happiness are mutually exclusive. However, Kant explicitly states that the distinction between the two ‘is not … an opposition’, and that ‘pure practical reason does not require that one should renounce claims to happiness’ (CPrR 5: 93). Rather, he wishes to claim that ‘as soon as duty is in question one should take no account of’ empirically conditioned practical reason. Kant is not saying that morality is opposed to happiness in the sense that to be moral is necessarily connected with unhappiness. If it were morally necessary to renounce happiness, then it would be counter to morality to make the pursuit of happiness an object of the will, and the pursuit of happiness would be a moral evil. But if that were the case, then happiness would not be a concern for practical reason in the sense required to get the dualism off the ground. So the claim that there is no such opposition between virtue and happiness does not tell against the case for a dualism; in fact it helps the argument, because there can only be a dualism if practical reason has a legitimate interest in promoting the ends of both happiness and virtue.

We get the kind of dualism we need, however, when we approach the question from the concrete perspective of an individual choosing in this world how she will determine her will. To make the categorical imperative the master maxim of the will is not tantamount to renouncing happiness; rather, it is to choose not to prioritize the pursuit of happiness over all other ends. But when we prioritize morality like this, we risk jeopardizing the achievement of happiness (G 4: 396). And those who are more prudent in their considerations do not share this risk.

We can show this using some basic decision theory. Prudence dictates that individuals maximize their self-interest. In a prisoner's dilemma, where both prisoners are innocent, the virtuous person will maintain her innocence, while the prudent person will maximize her advantage by confessing. Similarly, in situations where voluntary cooperation maintains a public good, the virtuous person will be willing to engage in the cooperative enterprise, and the prudent person will free ride when that maximizes her advantage. In general, when there are no enforceable sanctions against non-cooperation, the prudent person will be non-cooperative. However, no virtuous person will elect to be a free rider because free riding is based on a non-universalizable maxim. This suggests that there will be a rough proportion between being virtuous and receiving the ‘sucker's payoff’, and between acting prudently and maximizing non-moral goods. Kant's comment that the maxims of virtue and prudence ‘impose … severe limitations of one another, operating to each other's detriment’ (CPrR 5: 112) can be seen as recognition of this conclusion.

We now have the elements for the kind of dualism we need: the pursuit of happiness is not a moral evil, nor should we renounce it on behalf of virtue. Furthermore, reason has a ‘commission’ from our natural side to pursue happiness. Yet we have good reason to anticipate that the more we are virtuous, the less likely it is that we will be happy. Prudence therefore counsels against always prioritizing the moral law. But prudence is a form of reasoning, even if it is normatively trumped by the obligation to make the categorical imperative the supreme determining ground of the will. So, the opposition is not simply that, as a matter of fact, we would be better off in renouncing the moral law, but that the faculty of reason itself generates opposing moral and prudential imperatives. An agent's rational take on her situation is bifurcated into prudential and moral assessments, creating a dualism in the space of practical reason between prudence and virtue. Prudence maximizes the chances of happiness, virtue minimizes them; and the source of this bifurcation is the very faculty that is supposed to achieve the highest level of synthesis, namely, reason. Reason seeks a synthesis of both values in the proportionality thesis, but since we know that there is no necessary empirical synthesis of virtue and happiness, we are left with an object of hope. And since we know that adopting the maxims of virtue is a bad strategy for maximizing happiness, the ends of pure reason are confounded by experience. This motivates a sense of dissatisfaction on the part of pure reason when agents prosper beyond a level proportionate to their virtue. Thus we have a formal basis for the idea of just deserts.

The last step in deriving the proportionality thesis from Kant's conception of practical reason comes from a consideration of the role of reason. Reason is distinguished from the understanding in virtue of the fact that it seeks the highest and most all-encompassing synthesis in the unity of ideas (A299/B356; see also LR 28: 993). Practical reason seeks an all-embracing unity of its ends in the idea of the highest good (CPrR 5: 107–9). The proportionality thesis is the most complete synthesis of virtue and happiness because it allows us to articulate the moral law in terms of happiness: morality is the science that teaches us how we are to become worthy of happiness (TP 8: 278), and happiness is now understood from a moral perspective. Happiness is caused by virtue in the ideal of pure practical reason, but this causality is not based on the idea that the exercise of virtue can create the conditions for happiness. Rather, it is that the exercise of virtue ought to be rewarded by happiness, and the highest good is when this is the case. From the perspective of pure practical reason, happiness is deserved and not something we should count ourselves lucky to possess. The synthetic a priori of unity of virtue and happiness is achieved when each idea can be understood in terms of the other.

In contrast, Reath's maximization thesis envisions the two values maximized, as it were, side by side. In the ethical community, universal virtue leads to universal happiness, but Reath does not think of this as a conceptual synthesis of virtue and happiness. The exercise of virtue makes the maximization of happiness possible, but it does not create a moral necessity for the instantiation of happiness. From a maximization perspective, happiness is not understood as something moral beings deserve, and virtue is not understood as a disposition that deserves happiness.

6 As noted in section 3, Reath does not distinguish between actions and ends when making the futility argument, and it might be thought that a version of the futility argument could work if it focused on actions rather than ends. Instead of arguing that the proportionality thesis cannot be an object of the will because it is an unrealizable goal, it could be argued that it cannot be an object of the will because there is nothing we can do to instantiate it.

This is Jeffrie Murphy's argument (Murphy Reference Murphy1966) against John Silber's suggestion that we can promote the realization of the proportionality thesis by ‘rearing children, serving on juries, and grading papers’ (Silber Reference Silber1963: 183). Murphy argues that since we cannot infer the noumenal character of virtue from phenomenal behaviour, there is nothing we can do to promote the ideal of just deserts. Because we do not know what anyone deserves, there are no steps we can take towards approximating the highest good. Murphy goes on to argue that justice is a matter of ‘rewarding (or punishing) certain behaviour patterns in the hope of encouraging (or discouraging) their recurrence’ (Murphy Reference Murphy1966: 109).

If Murphy's argument stands then a version of the futility argument remains in effect, for we cannot promote an end if there is no action we can take towards its realization. The problem is that Murphy's argument undermines more than just Kant's doctrine of the highest good. Take criminal punishment: Kant thinks it is primarily justified as a proportional response to a prior act of wrongdoing (MM 6: 332–3). But a necessary condition for judging an action to be a ‘wrongdoing’ is that it is imputable to the autonomous choice of the agent (A448/B476).Footnote 6 We can only impute actions to agents if we believe that they are free to act independently of empirical determining grounds. Now autonomy is not an object of knowledge; nevertheless, Kant thinks that our ‘consciousness’ that we are subject to the moral law ‘forces itself upon us’ (CPrR 5: 31). Since ought implies can, this ‘fact of reason’ provides objectively sufficient grounds to believe that finite rational beings have autonomous wills (A820/B848).Footnote 7 The relevant – transcendental – conception of freedom is, therefore, a noumenal postulate; it is another idea of reason, but also an a priori representation. So the Kantian claim is that we universally and necessarily impute actions to the free choice of the will when we consider them under the guidance of reason (rather than the understanding). We simply cannot remain within the context of Kantian morality at all if we do not assent to the non-empirical claim that actions are imputable. Without this, autonomy is a fiction and virtuous willing itself is a futile endeavour. So Murphy's contention turns out to be a radical departure from Kant's practical philosophy.

Once we allow for the imputability of actions, we can make an inference about an agent's degree of viciousness, which allows us to use the idea of proportionality in moral judgement. If someone willingly commits a crime then we can say that they are at least as vicious as anyone willing to commit that crime. But we cannot infer from this the noumenal condition of the will. Agents may be more vicious than their actions indicate, but they cannot be any less so, if actions are imputable. In this light we can consider what kind of punishment matches the gravity of the crime, and trim our judgements to considerations of proportionality between vice and just desert. The realization that we do not have noumenal insight into the will creates an onus to temper judgement in case we punish more severely than is merited: Judges should be careful not to punish disproportionately based upon an intuitive assessment of a guilty party's character. Those who temper their judgements because of their lack of insight into character make an effort to instantiate proportionality. On the other hand, if a judge punishes a crime more severely than she believes is warranted, then she has intentionally chosen not to instantiate proportionality. If the proportionality thesis is the highest good, then this is a condemnable choice. But if there is a morally evaluable choice not to instantiate proportionality, then it must be possible to try to instantiate it, even if it cannot be perfectly instantiated. A visceral dislike of an individual's character could be a motive for ignoring these epistemological limitations, and for choosing not to approximate proportionality in punishment. But we can also choose to promote the realization of proportionality by putting our affective responses aside, and punishing on the basis of actions alone.Footnote 8

Lara Denis writes that ‘Kant does not dismiss phenomenal virtue as a mere pretender to true virtue; instead he describes it as virtue's “empirical character”, meaning that phenomenal virtue is the form in which true virtue appears to us’ (Denis Reference Denis2006: 513). I take it that Denis's point is that a person's virtue appears to us in a certain way, and this provokes character judgements. For example, when we see a customer treating a shop assistant badly for some trivial, unjustifiable reason, we might say that the customer ‘appears to be a rude person’. We cannot know if she really is rude in a noumenal sense; however, the appearance gives us a provisional reason to think she is, and repeat offences will offer further evidence. We also know that there might be mitigating circumstances, and by taking this into account to temper our judgements, we strive to instantiate proportionality.

The phenomena/noumena distinction does not prevent us from making inferences from the phenomenal to the noumenal, but it does frame the way in which we must conceive of the validity of these inferences. On the one hand, the inferences of reason are a priori, and therefore universal and necessary inferences sufficient to produce rational conviction. But since we cannot claim knowledge of these objects, it is appropriate to factor this limitation into our judgements, in pursuit of the ultimately unrealizable goal of instantiating the highest good.

7 It has now been established that there is something we can do to promote the realization of the proportionality thesis, but there is also a question about the relationship of this idea to the moral law, and more generally about its place in practical reason. Silber argues that the purpose of the highest good, understood as the proportionality thesis, is to provide concrete duties for an otherwise purely formal moral principle (Silber Reference Silber1959). Other commentators have pointed out, however, that there is no argument for a duty to promote the proportionality thesis (Beck Reference Beck1960: 244–5; Rawls Reference Rawls2000: 314). Kant's claim that the highest good is an a priori object of the will seems to support Silber's contention (CPrR 5: 113), but Kant also claims that the highest good answers the last of the three basic philosophical questions: What can I know? What ought I to do? What may I hope? Yet we would expect it to be an answer to the second question if it is a concrete duty. Now Kant writes that the first question is theoretical, the second practical, and the third both theoretical and practical (A805/B833). This gives us space to consider whether, as an object of hope, the highest good has a place in practical reason separate from the doctrine of quotidian duties. To explore this possibility, I will develop the claim that the highest good provides a teleological context for practical reasoning.

First, a formal analysis of practical reason can explain why the highest good is an a priori object of the will, but not a material duty. An object of the will is an inference from a practical deduction. Since Kant's model of deduction is the syllogism, there should be a general proposition as the major premise, and a minor premise identifying an instance that falls under the major premise (A330/B386–7). The major premise in pure practical reasoning is the moral law (CPrR 5: 90), and finite rational nature is a class of agents subject to this principle. Assuming this is a valid deduction then the conclusion follows necessarily from the premises. Therefore, the highest good is an a priori object of the moral will. Hence, we need not infer that the highest good is meant as a doctrine of material duties from the claim that it is a priori necessary to promote the highest good.

How then does this a priori object of the will figure in the practical reasoning of a finite rational agent? Kant thinks that pure and empirically conditioned practical reason are fundamentally distinct determining grounds from which the will derives objects (CPrR 5: 15). Conclusions from practical deductions (Wille) are objects of choice (Willkür) (MM 6: 213), the natures of which are determined by the premises of the deduction. The first formulation of the categorical imperative is the most general major premise in practical reason, as it is just the form of a universal practical proposition. Applying this to finite rational nature should produce a general object of the will. The highest good meets this expectation since it is the representation of the unconditioned end of finite rational willing. It is a practical doctrine insofar as it is an end, but it also raises theoretical issues because it is a representation of the unconditioned. Since it is an a priori object of practical reason, it is a necessary and universal object of the will, but its possibility cannot be cognized from the perspective of the understanding. However, transcendental idealism establishes the limitations of this perspective, making it dogmatic to conclude that the highest good is an impossible goal. We can thus clear epistemological ground that allows us to hope for the realization of the highest good.

What then are we supposed to do in the light of this object of hope? First, we must remember that, just as the moral law is the most general practical principle, so the inference to the highest good is the most general practical syllogism for finite rational agents. As more or less specific categorical imperatives can be articulated, so more or less specific practical syllogisms can be constructed; and as concrete duties are grounded in the formula of universal law, the promotion of the highest good is the ultimate purpose of objects deduced from specific categorical imperatives. Thus the highest good provides the teleological context within which we understand our morally motivated activity.

With regard to this context, Kant writes that ‘it cannot possibly be a matter of indifference to reason how to answer the question, What is then the result of this right conduct of ours? nor to what we are to direct our doings and nondoings’ (R 6: 5). Two reasons come to mind why this might be so. First, agents who do not have a conception of the ultimate end of their actions can come to believe that there is no purpose to their morally motivated efforts. They would not be thereby justified in renouncing the moral vocation, but this anxiety has the psychological potential to undermine their moral resolve. Kant considers a case like this in his discussion of the ‘Spinozist’ (CJ 5: 452).

Second, conceiving of the ends of action helps us to effectively direct our efforts towards their realization. Now, we synthesize the pursuit of happiness with moral obligation by making ourselves worthy of happiness; and we make the pursuit of happiness an object of approbation in the eyes of an ‘impartial rational spectator’ (G 4: 393) if we promote the morally permissible happiness of others (CPrR 5: 34).Footnote 9 Since thinking about ends leads us to consider how to realize them, the idea of the highest good should influence this deliberation. Drawing an analogy with prudence helps to show how this might work.

Happiness is the ideal synthesis of our natural ends, and the highest good is a synthesis of our moral ends. Agents without a concept of happiness are heedless of how their choices will provide for the maximal satisfaction of their desires. They might be lucky, but they are imprudent, because their choices are determined by immediate inclinations. Prudence requires the ordering of desires, and often foregoing some of them. Similarly we can exercise ‘moral prudence’ to best promote the permissible happiness of others. For example, it might be better for an aid-worker to refrain from criticizing a corrupt government in order to ensure access to a population in need. We might also exercise moral prudence by identifying ways to simultaneously discharge our duties to ourselves and to others. For example, an agent with a talent for group organization might help others more effectively by using this talent to better organize the structure of an aid agency, rather than by working directly with people in need.

We can think of the highest good, then, as the answer to questions that specifically ask about the purposes of actual or proposed courses of action. But the attempt to get clear on what it is we are hoping to achieve when we act has the potential to influence the way we execute a policy. As we have a duty to promote the permissible happiness of others, we must become concerned with practical questions regarding the means to this end. But we also must remember that we cannot fully realize this goal. There is the potential to approximate the ideal, but since we cannot be sure of the outcomes of our efforts, we can only hope that our actions will be to good effect.Footnote 10

8 I have argued that the proportionality thesis is Kant's conception of the highest good, and discussed how it fits into the wider context of his practical philosophy. Nevertheless, Kant does sometimes refer to the maximization thesis, and as shown in section 2, the two theses generate opposing evaluations of a universal hell and a completely unjust world. So if we want to save Kant's doctrine from inconsistency, we must see if the two versions of the idea can be reconciled. That this is possible is already suggested by the idea of a world that is maximally virtuous and happy. In this world, both values are in proportion to each other. We saw in section 2 that we cannot adopt this idea as a standard of evaluation because it does not distinguish between the proportionality and maximization theses; however, we can fit this idea into the theory when we consider it from the perspective of agency.

Since virtue is unconditionally good, the proportionality thesis includes the moral demand that we, as agents, must strive to be virtuous. Now, the maximization of virtue in an ideal world would cause the maximization of happiness; therefore, it is consistent for Kant to think of the highest good as the maximization of happiness and virtue. The maximization thesis is a corollary of the proportionality thesis when viewed from the perspective of goal-directed agency. We can see then that Kant not only needs to distinguish between the proportionality and maximization theses, he also needs to distinguish between the highest good understood as a principle of evaluation and as an object for agency. A universal hell fulfils the proportionality requirement for an impartial evaluation of a certain kind of situation, but it is clearly not the best world for which we could wish or will. From an evaluative perspective determined by the highest good, a completely unjust world is worse than a universal hell because it contains people who deserve but do not enjoy happiness. It represents the complete disappointment of morally inspired hope. Conversely, the misery of the inhabitants of a universal hell is not a matter for regret when considered from the same perspective because it is rational to hope that they fail in their bad intentions. There is, however, another possible evaluation, this time from the perspective of the moral law. In this case, a universal hell is more condemnable than the completely unjust world because it is universally vicious. Different judgements are possible based on different principles of evaluation, and it is necessary to make these distinctions if we are to develop a coherent picture of Kant's theory of the highest goodFootnote 11.

Abbreviations: Kant's Works