Kantian Eudaimonism

Abstract

My aim in this essay is to reorient our understanding of the Kantian ethical project, especially in relation to its assumed rivals. I do this by considering Kant's relation to eudaimonism, especially in its Aristotelian form. I argue for two points. First, once we understand what Kant and Aristotle mean by happiness, we can see that not only is it the case that, by Kant's lights, Aristotle is not a eudaimonist. We can also see that, by Aristotle's lights, Kant is a eudaimonist. Second, we can see that this agreement on eudaimonism actually reflects a deeper, more fundamental agreement on the nature of ethics as a distinctively practical philosophy. This is an important result, not just for the history of moral philosophy but for moral philosophy as well. For it suggests that both Kantians and Aristotelians may well have more argumentative resources available to them than is commonly thought.

All the main conceptions in the tradition of moral philosophy must be continually renewed: we must try to strengthen their formulation by noting the criticisms that are exchanged and by incorporating in each the advances of the others, so far as this is possible. In this endeavor the aim of those most attracted to a particular view should be not to confute but to perfect.

—John Rawls (1999b: 302)

The title of my essay, ‘Kantian Eudaimonism’, might seem surprising, even contradictory, for if there is anything that Immanuel Kant is not, it is a eudaimonist. Indeed, it seems clear enough that eudaimonism—understood as the view that moral actions are good or choiceworthy because they make their agent happy—is one of the main targets of Kant's criticism: something close to the very opposite of the true ethical theory, inverting the proper place of morality and happiness in a rationally ordered life. Thus, it is no surprise to find Kant, in his more historical moods, all but defining his position in the history of moral philosophy in terms of his rejection of eudaimonism. The ethical theories of the ancients and their Scholastic successors—extending, in crucial respects, through the German rationalist tradition—were eudaimonist through and through. So, the Kantian moment is, to borrow a phrase, one of standing athwart history, yelling ‘Stop!’—not, of course, in a reactionary sense, but in a revolutionary one. History was wrong and must be reformed; the old order swept away and a new one set down in its place.

Now, despite thinking that Kant is basically right about ethics, I think he is basically wrong about its history—and in ways that matter to our understanding of his moral philosophical project. The first and most obvious thing to say here is that it is far from clear that all, or most, or even many of the leading historical eudaimonists actually held the position Kant rejects. This is not to say that none of them did; perhaps Epicurus and his followers are examples. But it seems obvious that the founding fathers, so to speak, of eudaimonism—Socrates, Plato, and especially Aristotle—did not. Though Aristotle at least—I am just going to focus on him here—certainly accepts that moral actions are good or choiceworthy because they make their agent happy, he very clearly does not understand this relation in the same way that Kant does. In particular, he does not believe that moral actions are good because, say, they satisfy the agent's sensible desires—what Kant calls ‘inclinations’—or give rise to sensible pleasure—what Kant calls ‘agreeableness’. Perhaps moral actions do do this, but even if they do, it is not in virtue of their so doing that they are good. Consequently, it seems natural to conclude that despite what Kant thinks, he and Aristotle are in fact talking past one another. What he means by happiness and what Aristotle means by happiness are so different that any disagreement between them is merely apparent. They are simply on about different things.

I think this conclusion is basically right. But I do not want to stop there, for, despite Kant's own confusions about the history of ethics, there is a way of understanding his engagement with eudaimonism that brings into relief fundamental features of his own view, features that can be otherwise easy to overlook. In particular, there is a way of reading Kant that allows us to see him and Aristotle as not just talking to rather than past one another, but as actually agreeing on basic points. We can see this in two ways. First, once we approach the Aristo-Kantian encounter not from Kant's point of view but from Aristotle's, we can see that, while Aristotle is not a eudaimonist by Kant's lights, Kant is a eudaimonist by Aristotle's lights. And second, and more importantly, we can see that this agreement on eudaimonism actually reflects a deeper, more fundamental agreement on the nature of ethics as a distinctively practical philosophy.¹

My hope, then, is that by using Aristotle as a foil for Kant, we can better understand some of the rudiments of the Kantian position. I do not claim that everything that Kant says is consistent with the view I lay out, though I suspect many of the apparent inconsistencies are just that: apparent. I only claim that it captures a significant, deep-seated, and underappreciated strand in his thinking, one that provides the basis for a fruitful reconsideration of the Kantian approach to ethics, especially in relation to its assumed rivals.

As this suggests, I take my results to be important not just for the history of moral philosophy but also for moral philosophy itself, insofar as we are interested, as we should be, in clarifying the possibility space, so to speak. Is there as much fundamental disagreement in ethical theory as people think? How deep do these disagreements go? If it turns out that there is less disagreement or it is not as deep, might the philosophical resources of one view be brought to bear by another? Might this even lead to, of all things, progress? Alas, I am not able to address these further questions here, but if I am right about Aristotle and Kant, I hope to have made some preliminary steps toward a constructive answer.

I. Happiness

I begin with Aristotle. What is Aristotelian happiness (eudaimonia)? Recall how happiness enters the argument in the Nicomachean Ethics. It enters as an uncontroversial identification of the good that all rational activity seeks. The many and the wise agree, Aristotle says, that the good is happiness. They just disagree about what makes us happy; some say pleasure, some say honor, some say virtue, and so on. This makes the introduction of happiness seem trivial, and, in a way, I think it is. But it is not completely trivial, for it does add something to say that the good is happiness. In particular, it adds that the good is living well and doing well, since, as Aristotle tells us, ‘living well and doing well are the same as being happy’ (Aristotle 1999: 3 [1095a18–19]). So understood, living well and doing well are not candidates for what makes us happy, to be placed alongside pleasure, honor, and the like. They are explications of what happiness is, refining the question rather than delimiting its answer.

To represent this difference, let us say that Aristotle, implicitly at least, distinguishes between a concept of happiness—which is the same as living well and doing well—and the various conceptions of happiness—the various determinate accounts of what it is to live well and do well: pleasure, honor, and the like. Aristotle has his own favored conception, of course, but he recognizes that he must give an argument for this. Identifying the correct concept of happiness, he thinks, is easy. Identifying the correct conception of happiness is hard—or, at least, harder.

Kant identifies happiness (Glückseligkeit) in a number of different ways, not all of which are obviously equivalent. But Kant's dominant way of identifying happiness is something like this: as the sum of the satisfaction of inclination—sensible desire—or perhaps the feeling of agreeableness—sensible pleasure—to which such satisfaction gives rise. So understood, happiness is, as Kant says, an indeterminate idea. Though we all seek happiness, the happiness we seek will differ because of our different inclinations and feelings. If you want or enjoy sunsets in serene settings, your happiness will look one way; if you want or enjoy violent video games, it will look another. This indeterminacy allows us to attribute to Kant a similar structure to that we attributed to Aristotle. There is a concept of happiness—the sum of the satisfaction of inclinations or the agreeableness to which such satisfaction gives rise—and there are various conceptions of happiness—the sum of the satisfaction of a determinate set of inclinations or the agreeableness to which such satisfaction gives rise. And while Kant apparently thinks that identifying the concept of happiness is easy, identifying any individual's conception of happiness is hard, since, as he puts it, an agent ‘can still never say determinately and consistently with himself what he really wishes and wills’ (Kant 1997b: 28 [G 4:418]). That is, he cannot say, certainly not in advance of experience, what in fact will make him happy.

Suppose this is right. How, then, are we to make sense of Aristotle's and Kant's differing views of happiness? Given what I have just said, there seem to be two possibilities: first, they share a concept of happiness but differ as to their conceptions; or second, they differ as to the concept itself, rendering questions about conceptions moot. Put this way, the answer should be obvious. They differ as to the concept and not merely as to the conception.

Again, for Aristotle, the concept of happiness is living well and doing well. For Kant, by contrast, the concept of happiness is the sum of the satisfaction of inclinations or the agreeableness to which such satisfaction gives rise. But these, it should be clear, are very different concepts. From Aristotle's point of view, Kant's concept of happiness at best identifies one possible—and not very plausible—conception of happiness, something akin to the appetitive life of the hedonist. From Kant's point of view, Aristotle's concept of happiness is no concept of happiness at all, since it bears no obvious, much less necessary, connection to inclination or agreeableness, and so runs afoul of Kant's claim that ‘all elements that belong to the concept of happiness are without exception empirical’ (Kant 1997b: 28 [G 4:418]). Hence, as a conceptual matter at least, Aristotle could allow for a happiness that consists in purely intellectual activity. Kant, by contrast, cannot. A happiness ‘not based on empirical causes’, he is clear, would be a ‘self-contradictory absurdity’, the ethical equivalent of a round square (Kant 1996: 142 [MS 6:377]).

If this is right, then it should be obvious that though Aristotle and Kant both employ terms naturally translated as happiness, they use these terms in such different ways that they are best seen as talking past one another. Since they have different concepts and not merely different conceptions of happiness, the disagreement between them is merely apparent. They are simply on about different things.

II. Well-Being and Well-Doing

Now, one might be inclined to leave the issue there. But there is more to say. In particular, Kant gives us a way of thinking about this conceptual difference that allows for more similarity than meets the eye. For the Aristotelian concept of happiness is not absent from Kantian theory. It just does not appear under the name of happiness.

Although, as I have said, Kant's leading account of happiness is in terms of the satisfaction of inclination or the feeling of agreeableness to which such satisfaction gives rise, he does often characterize happiness in apparently more generic terms as ‘a complete well-being’ (Kant 1997b: 7 [G 4:393]) or a ‘maximum of well-being in my present condition and in every future condition’ (Kant 1997b: 28 [G 4:418]). One might think that this identification of happiness and well-being is, in effect, trivial; that for Kant at least it adds nothing to say that someone who is maximally happy also enjoys a maximum of well-being. But I do not think this is quite right. For when one looks, especially, at Kant's discussion of well-being in the second Critique, it becomes clear that Kant has something very particular in mind when he talks about well-being.

Consider, for example, his discussion of the contrast between well-being and good in the chapter ‘On the Concept of an Object of Pure Practical Reason’. Well-being (or ill-being), he says, ‘always signifies only a reference to our state of agreeableness or disagreeableness, of gratification or pain, and if we desire or avoid an object on this account we do so only insofar as it is referred to our sensibility and the pleasure and pain that it causes’ (Kant 1997a: 52 [KpV 5:60]). The main thing to note here is well-being's essential connection to sensible feeling, to agreeableness. Given this claim, it should be no surprise that Kant connects well-being and happiness, which, in the second Critique, is itself introduced in terms of ‘a rational being's consciousness of the agreeableness of life uninterruptedly accompanying his whole existence’ (Kant 1997a: 20 [KpV 5:22]). If well-being always signifies a reference to our state of agreeableness, and happiness is a maximum of agreeableness, then of course happiness is a maximum of well-being.

But Kant also connects well-being, and so happiness, in these passages to another concept: the worth of one's condition. By this, I take it that Kant means to allude both to an individual's state of agreeableness, and more broadly, to the idea of how things are going for an individual, where the ‘going’ is really an under-going, something of which the individual is the patient. The worth of one's condition, then, is a function of what happens to an individual, and in particular, how events in the world (including her actions) affect her feelings—which is why Kant connects it explicitly to sensibility, which is a receptive or passive faculty.

So understood, well-being contrasts with good. Where well-being refers to a state of agreeableness, good, Kant tells us, refers first to actions and then to persons as the agents of those actions. Where well-being is tied to sensibility and feeling, good is tied to will and reason. And where well-being marks the standard for the worth of one's condition, good marks the standard for the worth of one's person—where the worth of one's person is not a matter of how things are going for an individual but a matter of how an individual is doing: a function not of what happens to her but (literally) of what she does, not of her passivity but of her activity. On the basis of this last point, we can say, as Kant sometimes does say, that we should distinguish well-being (Wohlbefinden) not simply from good but also, and equivalently, from well-doing (Wohlverhalten).

With this distinction made, consider the aforementioned differences between Aristotle and Kant on happiness. Recall, Aristotle's concept of happiness is living well and doing well, and Kant's concept of happiness is the sum of the satisfaction of inclinations or the agreeableness to which such satisfaction gives rise. Given the present discussion, however, we can understand this conceptual difference in a new light. For, while Kant's concept of happiness is a well-being concept rather than a well-doing concept, Aristotle's concept of happiness is a well-doing concept rather than a well-being concept.

Aristotle and Kant, then, differ not only in their concepts of happiness. They also differ in the kind of concept their concept of happiness is, and so the role that happiness, so understood, should play in ethical theory. Moreover, we can also now see that not only are Aristotle and Kant talking past one another when they are talking about happiness, though they are. We can also see that they are talking to one another, though not when they are talking about happiness. For Kant, of course, is, in the present terms, not simply interested in well-being, he is also (and primarily) interested in well-doing. But his account of well-doing is not given by his account of happiness. Rather, it is given by his account of morality. So, putting the pieces together, while Aristotle and Kant are talking about different things when they are talking about happiness, they are talking about the same thing when Aristotle is talking about happiness and when Kant is talking about morality.²

III. Practical Philosophy

Now, this may seem a surprising conclusion, but I do not think it should be. For it reflects a rather deep agreement between Aristotle and Kant regarding the essentially practical nature of ethics, an agreement that has been largely obscured by the merely apparent disagreement about the role of happiness in ethical theory.

We can get at this by reflection on Kant's and Aristotle's identification of ethics as a kind of practical philosophy. That Kant identifies ethics as practical philosophy is obvious—so obvious that we rarely stop to consider the significance of the identification. At first blush, we may take Kant to be doing no more than signaling a primary concern with action. To this extent, he is implicitly rejecting a kind of Moorean vision of ethics as a ‘general enquiry into what is good’ (Moore 1993: 54). Kant's inquiry, by contrast, is not general but specific. He wants to know not about goodness as such, but goodness in action.

A brief comment on goodness. There remains a reluctance in some quarters to see Kant as concerned, at least in the first instance, with goodness at all. What is fundamental to his theory, some say, is deontic notions, for example, how one ought to act, rather than evaluative ones, for example, how it is good to act. But this is just a mistake, since, on Kant's view, ought is simply the guise under which imperfectly rational beings represent the good. As Kant says in Groundwork of the Metaphysics of Morals, ‘every practical law represents a possible action as good’ (Kant 1997b: 25 [G 4:414]), and his exposition of the famous distinction between hypothetical and categorical imperatives is explicitly in terms of the conditions under which an action is represented as good. All imperatives, no matter their kind, say that to do (or omit) something would be good. The specific difference between hypothetical and categorical imperatives, then, simply concerns what feature of the doing (or omitting) the imperative points to as good. As Kant says, ‘if the action would be good merely as a means to something else the imperative is hypothetical; if the action is represented as in itself good . . . then it is categorical’ (Kant 1997b: 25 [G 4:414]). Thus, we might equally call practical laws, as Kant himself does in these same passages, ‘laws (of the good)’ (Kant 1997b: 25 [G 4:414]). Ethics, then, as an inquiry—indeed, Kant thinks, a science—of action and its laws, is no more and no less than a science of the good.

Again, ethics is practical philosophy insofar as it is concerned with good action. But Kant actually has something more specific in mind. He is most explicit about this in Section I of the introduction to the third Critique, where he distinguishes theoretical and practical philosophies. He says there, unsurprisingly, that practical philosophy is concerned with action and its principles. But he goes on to make clear, in a way that he did not quite do so in earlier works, such as Groundwork of the Metaphysics of Morals, that it is not enough to mark the distinction between the theoretical and the practical to note that the latter concerns action and its principles while the former does not. For, he now claims, there can be principles of action that properly belong to theoretical rather than to practical philosophy: principles of action that concern the possibility of things in accordance with nature, notably, with the production of natural effects. That they are practical indicates something special about the cause of these effects—namely, that the causes are beings with the capacity to act in accordance with such principles—that is, they are rational beings, beings with wills—but the difference is not deep.

What does Kant have in mind here? Things like this: in order to bisect an angle, you should do this with your compass and do that with your straightedge; in order to shoot a projectile that will hit a target at such and such a distance, you should launch it at this angle with that force. In general, Kant thinks, whenever the rule in question concerns simply how to produce some effect, the rule is only superficially practical; it is, as he says in the first draft of this passage, practical in its form but not in its content (Kant 2000b: 4 [KUEE 20:197]). Such practical propositions, Kant says, are mere ‘corollaries of theoretical philosophy’—practical addenda to the theoretical sciences of geometry and physics, as in the above examples—and do not constitute a practical philosophy proper (Kant 2000a: 60 [KU 5:172]).

If there is a special branch of philosophy concerned with a distinctive (and distinctively practical) subject matter, then it must concern action not merely as productive of some natural effect but action considered in some other way. Which way is this? Kant's official answer here turns on a distinction between concepts of nature and concepts of freedom. However, explicating this distinction would take us too far afield, as well as miring us in some of the most difficult issues in Kantian philosophy. Luckily, then, Kant also puts the same point in other terms, ones that are more in line with our current concerns. For he says that superficially practical rules are only ‘technically practical’ and the properly practical rules are ‘morally practical’ (Kant 2000a: 60 [KU 5:172]). In putting the point this way, Kant is clearly adverting to the familiar discussion of imperatives to which I allude above.

Technically practical rules are hypothetical imperatives, rules that say that to do something would be good as a means. Morally practical rules, by contrast, are categorical imperatives, rules that say that to do something would good, not merely as a means but also as an end, in itself. Thus, what Kant seems to be saying in the third Critique is that the superficially practical rules—the ones that really belong to theoretical philosophy—concern actions that are good merely as means—and the properly practical rules—the ones that really belong to practical philosophy—concern actions that are good also as ends.

Here, then, is a key to solving the puzzle. I am looking for a way of considering action not merely as productive of some natural effect, for only such action could figure in the subject matter of properly practical philosophy. But if, as seems plausible, considering action merely as productive of natural effects is tantamount to considering the action merely as a means to the end that is that effect, then I have found what I am looking for. For the alternative is now obvious. Do not consider the action merely as a means. Consider the action also as an end. That is to say, the kind of action that is the concern of properly practical philosophy just is action that is its own end.

Now, it may seem a bit odd to think of this distinction in the value of action—instrumental versus final—in terms of a distinction between two different ways of thinking about action itself. But Kant is here aligning himself with a long tradition in moral philosophy, going back at least to Aristotle. For Aristotle, like Kant, as I have just explained him, distinguishes two senses of action. In Aristotle's terms: production (poiêsis) and action in a more restricted sense (praxis).³ (For clarity's sake, I call action in the restricted sense praction.) So, what is the difference between production and praction? Pretty much what we should expect. As Aristotle says, ‘production has its end in something other than itself, but action [that is, praction] does not, since its end is acting well (eupraxia) itself’ (Aristotle 1999: 89 [1140b7–8]). In other words, productions are actions whose end and so good is the effect of the action, its product; for example, housing building is a productive action, whose end, and so good, is the house. But practions are actions whose end and so good is itself; for example, virtuous actions, according to Aristotle, are practions. They have products, of course, but they are nonetheless good in themselves insofar as they are what acting well (in the relevant circumstance) consists in.

If this is right, then Kant and Aristotle agree that there are two ways of thinking about action. Moreover, they also agree that this distinction between kinds of action gives rise to a corresponding distinction between kinds of knowledge. To know how to produce is to have, in Aristotle's terms, craft knowledge (technê). By contrast, to know how to pract is to have, in Aristotle's term, practical wisdom (phronêsis). And Aristotle is clear: it is the latter rather than the former that we properly seek in practical philosophy. Again, Kant agrees, noting, for example, in the Metaphysics of Morals that practical philosophy, properly so called, is not a technically practical doctrine. On the contrary, it is ‘none other than moral wisdom’ (Kant 1996: 11 [MS 6:217]).

Again, Kant claims that not all rules concerning actions are properly practical—in particular, the ones that concern, in Aristotelian terms, productions are not. So which ones are? The ones that concern practions. That is to say, the proper subject matter or object of practical philosophy is action that is its own end and so good, because simply in virtue of acting in that way, the agent acts well. That is what ethics, practical philosophy, is really about for both Kant and Aristotle.⁴

IV. Eudaimonism

The Kantian concept of happiness is the sum of the satisfaction of inclinations or the agreeableness to which such satisfaction gives rise. Hence, it should be no surprise that Kant explicitly includes among the superficially practical principles, the ones that are mere corollaries of theoretical philosophy, ‘the general doctrine of happiness’ (Kant 2000a: 60 [KU 5:173]). For if happiness is as Kant takes it to be, then it should be obvious that principles that tell us how to be happy are principles that concern productions rather than practions. For the actions at issue here are actions that are good only as means to satisfying inclinations or arousing agreeableness. This means that Kant's deepest objection to eudaimonism, as he understands it, is not merely, as he often emphasizes, that it can furnish no laws: that there is no single thing or set of things that we must do in order to be happy, since the content of happiness is given by inclinations. For even if happiness were determinate—even if we had all the same inclinations and so there were a single thing or set of things we must do in order to be happy—Kant would still object. For it would remain the case that moral actions would be mere productions. But, again, ethics, as practical philosophy, is not concerned with productions. It is concerned with practions. Thus, given how Kant thinks about happiness, his deepest objection to founding ethics on happiness is that it is, in a way, self-contradictory. An ethics of happiness is, strictly speaking, no ethics at all.

But, again, if we think about happiness in an Aristotelian way, things look quite different indeed. For the Aristotelian concept of happiness is living well and doing well, where it is clear that the doing at issue is not just any old action but praction. What the happiness seeker seeks, then, is not simply to act to produce some effect, on the world or on herself. What she seeks, according to Aristotle, is to pract, to undertake action that is its own end and so good, insofar as that action is constitutive of acting well.

Figuring out which actions these are is the task of practical philosophy, the attainment of practical wisdom. And so, for Aristotle, unlike for Kant, there is nothing self-contradictory about founding ethics on happiness. On the contrary, it is (almost) tautologous. For asking how to pract is the same as asking how to live well and do well and thus how to be happy.

So, again, we can see that despite the confusions that naturally arise from their use of different philosophical idioms, Kant and Aristotle are in fact in deep agreement. While Aristotle is not, by Kant's lights, a eudaimonist, Kant is, by Aristotle's lights, a eudaimonist, since being a eudaimonist is (more or less) the same thing as being a practical philosopher.

V. Objections and Replies

To be clear, I am arguing that Kant and Aristotle are eudaimonists and practical philosophers to the extent that they center ethical reflection on how to act—more specifically, pract—well. I am not arguing that they agree on any particular account of what so acting involves. That is to say, in the terms I introduced above, I am claiming that Kant and Aristotle share a concept of the object of practical philosophy, acting well, but I am not claiming that they share a conception of it, a determinate view as to what acting well actually consists in.

This conceptual focus is important, since it explains why some of the most obvious objections to my conclusion miss their mark. For example, one might naturally react to my claim that Kant is, by Aristotle's light, a eudaimonist by pointing to the many notable elements of Aristotelian happiness, such as pleasure and external goods, that do not seem to feature in Kantian morality. But even if it were true that Kant and Aristotle differed in this way, it would not undermine my central claim. Why? Because the difference would not amount to a difference in the concept of the object of practical philosophy. Rather, it would amount to a difference in conception. Put another way, if I am right about Kant and Aristotle, then the disagreement between them, such as it is, should be understood in much the same way as the disagreement between, say, Aristotle and the Stoics; that is, as a disagreement between eudaimonists—about external goods, for example—rather than a disagreement between eudaimonism and its other.

Although arguing the point is beyond the scope of this essay, it is worth noting that it is not obvious that Kant and Aristotle do differ in this way. With respect to pleasure, we have gone wrong because we have assumed that pleasure, for Kant, is something sensible and so extrinsic to rational agency. But this is a mistake. Agreeableness, the pleasure that figures in Kant's concept of happiness, is so extrinsic. But that is not the only kind of pleasure he admits. There is also a sui generis intellectual pleasure that is intrinsic to rational agency, marking excellent exercises of it. Even for Kant, then, acting well feels good.⁵

The external goods are harder, in part because it is harder to know exactly why Aristotle thinks we need them. Suffice it to say, though, at least some of Aristotle's thoughts about the importance of such goods seem available to Kant too—for example, his claim that ‘happiness needs external goods to be added, as we said, since we cannot, or cannot easily, do fine actions if we lack the resources’ (Aristotle 1999: 11 [1099a30-1099b1]). This seems consonant with Kant's thought that, while good willing does not require productive success, it does nonetheless require ‘the summoning of means’ (Kant 1997b: 8 [G 4:394]). Indeed, how could it not, if the will is by nature a causal power, in which volition of an end constitutively involves the use of means? It would seem to follow, then, that without adequate means one could not will at all, and so, a fortiori, could not will well.

Still, though, one might object that even the limited agreement I have emphasized—concerning concept rather than conception—goes too far. For, one might say, even if Kant and Aristotle are primarily concerned with acting well, the fact that Aristotle represents this concern under the guise of happiness and Kant under the guise of morality still matters. After all, it seems, Aristotelian happiness remains a prudential concept. Thus, to say that acting well makes us happy is to say that acting well is (non-instrumentally) good for the agent: that it benefits her, is in her interest, and so on. But Kant, of course, would not agree. Acting well is good, he thinks, but not in this way. So, their shared interest in acting well notwithstanding, there remains an important sense in which Aristotle is and Kant is not a eudaimonist.

True, Kant does not think that moral actions are good for their agent, in the sense of benefit. Or, at least, he does not think that such goodness is the primary reason we should perform them. But neither, as far as I can see, does Aristotle.⁶

Again, Aristotelian happiness is (conceptually) living well and doing well. Of course, one might ask, whose living and whose doing? And the answer, obviously, will be the agent's. But I do not see how this brings in anything prudential. For, when pitched at this level of abstraction, determining how to be happy is no different than determining how one should live or what one should do. If there is an application of good for, it is not one of benefit but of rational agency itself: what it is good for an agent to do—that is, what she must do in order to count as doing well. In this sense, it could be that what it is good for an agent to do, as determined by practical reason, is to benefit herself; and surely sometimes it is. But it could equally well be that what it is good for an agent to do, as determined by practical reason, is to benefit others; and surely sometimes it is. As far as I can see, then, Aristotelian happiness is not a prudential good, in any interesting sense. For again, to say that moral actions make their agent happy is just to say that they make their agent successful qua agent; that by so acting, she is acting well. And with that, Kant would certainly agree.

But surely, one might insist, there must be something un-Kantian in centering ethical reflection on what is good, even if goodness is understood in a generic agential rather than specifically prudential sense. Indeed, this might seem to be the point of Kant's so-called ‘paradox of method’, which appears immediately after Kant distinguishes well-being and good in the way I describe above (Kant 1997a: 54 [KpV 5:63]). After all, he does say there that the good does not determine the (moral) law; on the contrary, the (moral) law determines the good.

True, Kant thinks that the law must determine the good. But as far as I can see, this is not at odds with my proposal. The first thing to say is, again, that my argument concerns the concept of the object of practical philosophy. It does not concern the conception. So, at this stage, the good is undetermined, and necessarily so. We know we seek the good, doing well. But we do not yet know, practically speaking, what this comes to—that is, what we must do in order to do well. How, then, are we to figure this out? By means of the law. But what does that mean? Something like this.

It is important to remember here that Kant's leading thesis about the law, argued at length prior to the paradox, is that it is a constitutive principle of rational nature, that is, a principle that expresses the kind of thing a rational agent is and so the kinds of actions of which a rational agent is capable. But if so, then the claim that the law determines the good is tantamount to the claim that our nature determines our good; what we are determines what we should do. But understood in this way, Aristotle would of course have no objection. For he, too, believes that the nature of a thing, which can equally be captured by its principle (archê), determines its good, what counts as success or doing well for that thing in general. (Hence, the function argument.) So understood, then, the paradox does not mark a point of difference between Kant and Aristotle. On the contrary, it marks a point of profound agreement about how to move from concept to conception in practical philosophy.

Again, this does not mean that Kant and Aristotle agree on what this move gets us. More specifically, since they do not fully agree on what we are, they do not fully agree on what we should do. It is here that Kant's emphasis on rational nature and Aristotle's emphasis on human nature becomes important, as does their more detailed accounts of what the specifically human form of rationality consists in. But that is not to the present point, which is just that these disagreements, such as they are, should be understood as intramural: a dispute, so to speak, among friends, who share a common view of what ethics is and so how ethical argument should be conducted.

VI. Moral Theory

Kant is not, then, the revolutionary he takes himself to be, for he is in fact executing, in his own way, a program with ancient antecedents. One might think that this diminishes Kant's position in the history of moral philosophy. On the contrary, I think it enhances it, not just because the ancient program is pretty plausible—more plausible, to my mind, than what has traditionally passed as the Kantian program—but because it allows for a more constructive engagement between Kant and other philosophers. Which brings me back to where I began. For though my primary interest here has been to understand the Kantian position, in its essentials, I also hope to have contributed, in some small way, to what John Rawls calls ‘moral theory’: an investigation into ‘the systematic and comparative structure of moral conceptions, starting with those which historically and by current estimation seem to be most important’ (Rawls 1999a: 341). Moral theory, so understood, is but one part of moral philosophy, but an important one, since it clarifies the possibility space in which moral philosophers work, and in particular, the fundamental choice points at or on which philosophers may divide.⁷

Because Aristotelian and Kantian ethics, by most any measure, meet the condition of being by historical and current estimation important, sorting out their relation—especially if that relation is, deep down, something approaching identity—would be of great significance, not just as a matter of historical understanding but philosophically too. For by revealing the conceptual infrastructure of each, one gains a sense of their essential load-bearing elements, so to speak, and so what pieces may be removed without serious damage—a rotten beam here—or added for serious improvement—a retaining wall there. Put less metaphorically, by thinking through these theories together, we allow for the possibility that Kantians could avail themselves of what have been thought to be Aristotelian arguments and Aristotelians could avail themselves of what have been thought to be Kantian ones, allowing us to address old problems in new ways and perhaps even to make progress in philosophy. I have not argued that this possibility is actual, but I hope to have made the case that it is worthy of further investigation. In this way, we may heed Rawls's counsel, articulated in the epigraph to this essay, aiming always in our arguments not simply to confute but to perfect.