Eudaimonism, ‘Egoism,’ and Degrees of Finality in Aristotle’s Ethics

And now for something completely different…

In some circles, the idea that Aristotle is an egoist passes for a truism. After all, he claims that the ultimate aim of human action is eudaimonia, happiness or flourishing; everything else is worthy of choice for its sake, while it — and only it — is worthy of choice entirely for its own sake and not for the sake of anything else. Could we possibly find a clearer statement of egoism than that? Yet in contemporary Anglo-American scholarship, the notion that Aristotle is any kind of egoist is extremely controversial. It is agreed that Aristotle, like most or all other philosophers in the ancient Greek tradition, is a eudaimonist, and hence that the flourishing or well-being of human beings has a central role to play in his thinking. But probably most Aristotle scholars would reject the suggestion that he is, at least without qualification, an egoist. To some extent, the dispute is terminological: ‘egoism’ brings a great deal of baggage with it, and in any case tends to suggest that a person should not only make his own well-being or flourishing the guiding aim of his life, but, further, that he should regard the well-being of others as a merely instrumental means to his own. It is widely, if not almost universally, agreed that Aristotle rejects that sort of purely instrumental view of regard for others, and many would take that rejection as decisive grounds for denying that Aristotle is an egoist. Predictably, however, matters are not so simple.

The term ‘egoism’ is used in a bewildering variety of ways, and I have no interest in shaping the way that others use it. For my purposes, though, some standard distinctions may be in order. There are at least three different claims that go under the name of ‘egoism’: psychological egoism, ethical egoism, and rational egoism. Each of these is supposed to be a kind of egoism because it gives an important explanatory role to an individual human agent’s self-interest. But they are claims about different things. Psychological egoism is a claim about human psychology: it is the claim that we always act for the sake of what we believe to be in our self-interest. Ethical egoism is a claim about ethics: it is the claim that we should always act for the sake of what truly is in our self-interest and never act contrary to it. Rational egoism, in the words of David Brink, “is a theory about the grounds of reason for action: it says that an agent has reason to do x just in case, and in so far as, x promotes his own interest, welfare, or happiness” (David O. Brink, Moral Realism and the Foundations of Ethics, Cornell 1989, 67).

It is important to see how rational egoism is distinct from both psychological and ethical egoism. Rational egoism differs from psychological egoism because it is a claim about normative reasons, or what we might call “good reasons,” as opposed to “explanatory” or “motivating” reasons; that is, it purports to tell us what makes it the case that we have good reason to do something, but it does not imply any particular claims about what kinds of reasons actually motivate us. Rational egoism is distinct from ethical egoism, as I have formulated it here, because ethical egoism tells us what we should do, but says nothing about why we should do it; it would be consistent with my formulation of ethical egoism that we should always act for the sake of what is truly in our self-interest because by doing so we will promote the greatest happiness of the greatest number, because God commands it, or because we just happen to feel like it. Conversely, rational egoism makes no claims about the content of well-being or self-interest, and in particular it does not entail that we should never act to benefit others for their own sake or that we should adopt a merely instrumental conception of how our own well-being is related to the welfare of others. In other words, rational egoism is, in principle at least, compatible with an apparently very un-egoistic theory of what is good for an individual human being. To avoid this sort of confusion, some (including me) have opted to abandon the language of ‘egoism’ in favor of less misleading alternatives. One candidate I have offered, drawing on Mark Murphy’s Natural Law and Practical Rationality, is agent-relative welfarism: this is a theory about what gives human beings reason to do something; it is welfarist in explaining all non-derivative reasons in terms of well-being or flourishing, and agent-relative insofar as it makes the agent’s own welfare the source of his non-derivative reasons for action. Agent-relative welfarism is consistent with reductive, instrumentalist ethical egoism about the content of well-being and what we have reason to do, but it does not entail that kind of egoism, since it is also consistent with conceptions of well-being on which positive, non-instrumental other-regarding relations with others are an irreducible aspect of human flourishing. That is the kind of view that many believe that Aristotle holds.

Many, but by no means all. Such prominent and influential scholars as Terence Irwin and C.D.C. Reeve have defended such a view of Aristotle, but likewise distinguished philosophers such as Richard Kraut and Anthony Price have questioned it. Both of these latter two have recently criticized the agent-relative welfarist interpretation of Aristotle as it has been applied in constructive, first-order philosophy by Dan Russell. Both Kraut and Price challenge Russell’s “eudaimonism” in their reviews of his recent book, Happiness for Humans. Here is Kraut:

Russell embraces a certain form of normative egoism, according to which each individual’s deliberative reasoning should have as its single final goal the happiness (that is, well-being) of that individual…

This is perhaps best understood as a two-level theory: at the ground level, one devotes oneself to others and does not treat them as a mere means to one’s own happiness. But why should one take up this non-instrumental stance towards others? In asking and answering this question, we move up a level: these loving attitudes towards others stand in need of justification, and the justification consists in the fact that taking this stance towards others is what makes one’s own life happy.

The objection to be made to this is that it does not recognize others as ever constituting a direct source of reasons. Another’s well-being never provides me, all by itself, with a reason to assist him; it does so only if and because I benefit from caring about him. If someone sitting next to me is in great pain, and I can stop the pain simply by lifting the electrical wire that is causing the problem, am I justified in assisting him only if and because my doing so partly constitutes a happy life for me? If I could be just as happy walking away, would that justify my doing so?

And here is Price:

Russell assigns to eudaimonia this privilege: it is ‘the final end, a good that halts deliberation by being the final source of reasons for our other ends’ (23). Following Mark LeBar, he distinguishes two levels of reasons: there are reasons for acting in virtue of the ends one has, and reasons to have those ends in the first place (26). The second level relates to the eudaimonia of the agent, whereas the first level allows me to act for another’s sake; however, the second level has priority in that it makes external considerations first-level reasons for the agent. Russell is willing to write, ‘The point of devoting oneself to another is for the sake of giving oneself a good life’ (27).

While Russell can claim to be doing full justice to NE I.12, he seems to me to be slipping from a tenable position to an unsustainable one. When a man enters into a relationship or obligation, he may reasonably want to make sure that this will occupy ‘a place in his life that he can live with’ (26). Certainly we can’t expect of one another, and wouldn’t desire for our own children, that they should take on burdens that, whatever the benefit to others, will land them with a life that they find unlivable. And we might express this by saying that what might be a reason for action for more robust agents is not a reason for them. Yet we eviscerate virtue of content if we deny that the virtuous agent is aware of reasons for action relating to harms and benefits whose source has nothing to do with his own welfare, and which connect with ends that he does not ‘adopt … for the sake of a good life’ (26). We may still say, if we wish, that, to count as reasons for him, they need to be put through the filter not only of what he can do, but what he can live with; but we concede too much to self-interest if we offer a dispensation not only to the man who can’t cope, but to the one can do better for himself.

Price and Kraut both object to Russell’s view in purely philosophical terms, without much appeal to Aristotle’s texts. This is because Russell’s book is primarily a work in first-order philosophy, not an interpretation of Aristotle. But just as Russell takes his view to be at least largely true to Aristotle’s, Price and Kraut likewise take their criticisms to apply to Russell’s reading of Aristotle, as well. Both of them have written excellent books about Aristotle in which they largely defend Aristotelian theses as attractive and insightful. Yet they agree in rejecting the agent-relative welfarist conception of Aristotelian eudaimonism precisely because it grounds an agent’s reasons for action solely in his own well-being. The disagreement isn’t just a terminological dispute about how to use the ‘egoism’ label; it is a substantive difference between rival interpretations of Aristotle that embody competing conceptions of human well-being and practical reason.

I don’t want to focus on the details of Russell’s argument, in part because I haven’t yet studied it carefully and in part because I have other fish to fry. Suffice to say that, on the fundamental issue, I think that Russell is right and that Kraut and Price are wrong. I think it is far easier, however, to show that Aristotle’s eudaimonism is a variety of agent-relative welfarism (or rational egoism, if you prefer) than it would be to show that his view is true. What I’ll do, then, is consider a passage from the Nicomachean Ethics that has been read as inconsistent with the agent-relative view, and then show, by a consideration of Aristotle’s claims about what it is to be teleios — ‘complete,’ ‘final,’ ‘end-like,’ or ‘goalish’ — that Aristotle is committed to the agent-relative welfarist view and that the allegedly incompatible passage is, in fact, perfectly consistent with it.


In EN 1.2, Aristotle tells us that the goal of politics — which is what he takes himself to be doing in the Ethics as well as the Politics — is none other than the human good (1.2 1094b6-7). Politics, in the relevant sense, is the most authoritative or controlling science or capacity (1094a24-26). It is controlling (architechtonikēs) and authoritative (kuriōtatēs) in the sense that (a) it determines which subjects people in cities should study and how far they should study them (1094a27-b2, 5), and that (b) all other sciences and capacities are subordinate to it (1094b2-5). Because it bears this relationship to all the other sciences and capacites, its end embraces the ends of those others. It follows, says Aristotle, that the end of this most authoritative and controlling capacity is the human good. It follows, he suggests, because (gar 1094b7) “even if it is the same thing for an individual and for a city, the good of the city is nonetheless evidently greater and more complete (teleioteron, 1094b8) both to attain and to preserve” (1094b7-9). My colleague Don Morrison has drawn on this passage to argue that, for Aristotle, the good of a political community is the highest object of practical reason and the ultimate end of human action (see his ‘Politics as a Vocation, according to Aristotle’, History of Political Thought 22, 221-41 and his ‘The Common Good’, in The Cambridge Companion to Aristotle’s Politics). On this sort of view, eudaimonia retains its central place in Aristotle’s thought, and a virtuous person will aim to achieve it for himself, but he will also aim to achieve it for his city, and not simply because this somehow contributes to his own eudaimonia, but more fundamentally because the good of the city is a greater and more complete good.

I want to challenge this reading of EN 1.2 by considering EN 1.7 1097a30-b6, where Aristotle tells us what he means in describing one thing as “more complete” than another. John Cooper, among others, has recently stressed (‘Plato and Aristotle on Finality and Self-Sufficiency’, in Nature, Knowledge, and the Good) the connection between being complete (teleion) and being an end (telos), arguing that we should abandon “complete” as a translation in favor of “final.” Leaving aside the details of his interpretation of the passage from 1.7, he seems right about this. Though there may still be something to be said for “complete” as a translation, Aristotle leaves no doubt about the connection between being teleion and being a telos insofar as a telos is an object of choice. In the passage in question, he distinguishes several degrees or scales of finality:

1. That which is pursued for its own sake is more final than that which is pursued for the sake of something else.
2. That which is never chosen for the sake of something else is more final than those things which are chosen both for their own sake and for the sake of something else.
3. That which is always chosen for its own sake and never for the sake of something else is unqualifiedly final.

This analysis gives us three levels of completeness. If we allow that goods are objects of pursuit and choice, we get:

a. Goods that are always chosen for the sake of something else and never for their own sake
b. Goods that are chosen both for their own sake and for the sake of something else.
c. Goods that are chosen solely for their own sake and never for the sake of anything else.

By the argument that follows, there is only one good of the third and highest grade: eudaimonia. It is always chosen for its own sake and never for the sake of anything else. Other goods — honor, pleasure, intelligence, and every virtue — are chosen both for their own sake and for the sake of eudaimonia. We choose these goods for the sake of eudaimonia, “supposing that we will be happy through them” (1097b5), but nobody chooses happiness for the sake of these things “nor for the sake of anything else at all” (b6).

If we try to apply this schema of grades of finality to the claim in EN 1.2, we encounter a problem. What that text seems to say is that the good of the city is more final than the good of an individual. By the analysis of 1.7, the good of the city could only be more final than the good of an individual if the good of an individual is chosen for the sake of the good of the city. Since, by the argument that follows, eudaimonia is never chosen for the sake of something else, the good of an individual either cannot be chosen for the sake of something else or it cannot be eudaimonia. Despite considerable controversy over Aristotle’s theory of eudaimonia, there is no question that he takes it to belong to individual human beings. It is a practical activity of the rational soul in accordance with excellence; living happy lives is something that individual people do, however much they need to do it together with others. So it looks like the good of an individual cannot be chosen for the sake of anything else: an individual’s good is eudaimonia, and nobody chooses — nobody could coherently choose — eudaimonia for the sake of anything else.

So if the good of the city is to be “more final” or “more complete” than the good of an individual, it can’t be in the way that eudaimonia is more complete than honor or pleasure or intelligence or virtue.

Or can it? We might be able to defend the consistency of Morrison’s reading of 1.2 with the claims of 1.7 by considering alternative ways of understanding the latter. We could take these as claims about types of goods or tokens of those types. Interpreted in terms of token goods, we can make sense of the idea of choosing some good for its own sake and for the sake of other, distinct tokens of the same type of good. Suppose Alcibiades chooses some course of action — say, entering four chariots in the races at Olympia — for the sake of honor. He pursues this honor for its own sake, and not simply as a means to power or bodily pleasures or what not. But now suppose that he also pursues this honor — the honor he receives for having entered so many chariots in the games, and the honor he receives for winning the race — for the sake of additional future honor, as well. He is pleased to be honored for entering so many chariots in the games, and would choose this honor even if nothing else were to result (cf. 1097b3-4). Yet he also chooses this honor because, he thinks, it will lead to even greater honor in the future; this act will draw people’s attention to him and lead them to honor him for his future exploits, as well. This seems to be a case of one token of honor being chosen for itself and for another token of honor.

But this seems not to be the sort of case that Aristotle is discussing in 1.7. He seems interested instead in distinguishing among types of good in terms of degrees of finality or completeness. After all, there is nothing remarkable about choosing one instance of pleasure for its own sake and for the sake of additional instances of the same kind of pleasure later, or mutatis mutandis for virtue, knowledge, or honor. In these cases, the future instances of the same kind of good do not supply any additional explanatory value or add any distinct kind of reason for the choice of the first instance. Choosing honor for the sake of pleasure, or knowledge for the sake of honor, however, is a different sort of thing than choosing one instance of pleasure for the sake of future instances of pleasure. If Alcibiades pursues honor not only for its own sake but also because it will improve his chances of seducing the most beautiful boys of Greece, this tells us something more about why he is acting in the way that he is, because the pleasures of sex are providing him with further reasons for seeking honor, in addition to the intrinsic attractions of honor. Pleasure is capable of motivating the pursuit of honor in place of or in addition to the motivational pull that honor can exert on its own accord.

Suppose something like that is what Aristotle has in mind. Would it help us to reconcile 1.2 with the claim that nobody ever chooses or pursues eudaimonia for the sake of anything else? Read in this way, the claim is that eudaimonia is capable of motivating the pursuit of these other goods in addition to the pull that those other goods exert on their own accord, but that no other good is capable of motivating the pursuit of eudaimonia (and this because we can’t coherently suppose that living well is something we can choose for the sake of anything else). But this would not exclude choosing or pursuing eudaimonia for the sake of other instances of eudaimonia: this might be the interpersonal parallel to Alcibiades’ choice of this honor now for its own sake and for the sake of some more honor later. The denial that eudaimonia is chosen for the sake of anything else, then, would amount to the claim that no other type of good can explain or motivate the pursuit of eudaimonia, not that one cannot pursue one token of eudaimonia for its own sake and for the sake of other token instances of eudaimonia.

But the claim, construed this way in terms of types, seems true in the intrapersonal but false in the interpersonal case. For any given agent, the pursuit of honor or pleasure or knowledge or virtue does not seem capable of explaining the pursuit of eudaimonia; nobody can seek to live well in order to enjoy themselves or to understand the universe. They seek those things, rather, in order to live well through them, as Aristotle says. But now consider one agent A acting in relation to some second agent B. It does seem sensible to suppose that A can act for the sake of B’s happiness not because of any motivational pull that B’s happiness exerts on its own accord, but only for the sake of the honor that A stands to gain from others as a result of being the sort of person who seeks other people’s happiness, or only for the sake of the opportunity to seduce beautiful boys that he thinks will result from promoting the well-being of others. Stated less in terms of motivations and more of ends, it is possible to pursue other people’s happiness solely for the sake of something else. One’s own happiness, however, can never be a means to anything else, not because of any contingent features of human psychology, but as a matter of what a final end is.

Notice that this conclusion is not the same as the claim that it is never possible to seek other people’s happiness for its own sake, nor even that it is never possible to seek other people’s happiness for its own sake and not for the sake of anything else. Aristotle may think that we cannot seek other people’s happiness for its own sake and not for the sake of anything else without ceasing to be distinct rational agents; the closest thing we find in Aristotle to a description of someone who acts for the sake of another person’s good with nothing more than an incidental concern for his own welfare is in his description of slaves, whom he regards as functioning not as fully independent rational agents, but as something like physically separate, practical-metaphysical parts of their masters (Politics I.4-6). But that is a different matter. What my argument above shows, unless it’s mistaken, is that we simply can’t pursue our own eudaimonia for the sake of anything else. But we can pursue other people’s eudaimonia for the sake of other things, whether or not we pursue it for its own sake.

What are the implications of this for Aristotle’s discussion of the grades of finality? Evidently, that the unqualifiedly final eudaimonia discussed in EN 1.7 is the individual agent’s own eudaimonia. This is hardly surprising if Aristotle supposes, as seems sensible, that finality is an agent-relative property, as it must be if it is to retain the essential connection with ends that Cooper et al. see in it (ends are necessarily agent-relative in that every end is an end of or for some agent). So the good of the city cannot be more final than the good of an individual by the analysis of EN 1.7. That analysis makes eudaimonia final without qualification because it is never chosen for the sake of anything else. Only the good of the person doing the choosing meets that criterion as Aristotle lays it out. So it seems that the good of the city can’t even be equally final with the good of an individual. If the good of the city were to be more final than the good of the individual, the good of the individual would have to be chosen for its own sake and for the sake of something else, and hence would have to cease to be eudaimonia as Aristotle characterizes it.

There is, however, a perfectly natural alternative way to read the passage from 1.2. When Aristotle tells us that “the good of the city is evidently a greater and more complete good both to attain and to preserve,” we need not take him as asserting that the city’s good is greater and more complete than the individual’s without qualification. Rather, we may read the infinitives as limiting the scope of application of the adjectives: it is a greater and more complete good to attain and preserve. It is not that the city’s good is worthy of choice for its own sake independently of its good for the people who attain and preserve it; it is instead a greater and more complete thing to attain and preserve than the good of any single individual taken on his own would be. The contrast, that is, is not between an agent’s interest in his own welfare and his interest in his city’s welfare; it is instead between an agent aiming to benefit one person and an agent aiming to benefit an entire city. If he is thinking straight, an agent who knows how to benefit an entire city and a single individual will rightly choose to benefit the entire city, precisely because benefiting the whole city will do as much good as would benefitting the single individual (assuming that he is a part of the city) and a great deal more good on top of that. We need not read “more complete” here in the precise sense it bears in 1.7, but even if we do, the idea will be not that the city’s eudaimonia is more complete than an individual’s, but that obtaining and preserving the city’s good is more complete than obtaining and preserving the good of some single person. The sensible political expert will not see his own welfare as a mere means to or part of the city’s greater good, but he will rightly regard benefiting a whole city as more worth doing, because it does more good, and thereby more fully actualizes his capacity for virtuous rational action.

Granted that this is the sort of view Aristotle holds, is it right? Some people are inclined to object that even if it allows for, or even demands, genuine benevolence and respect for others, it nonetheless makes our other-regarding reasons too derivative. Surely, some will urge, we just owe it to people to respect and promote their good simply because they are human beings, or persons, or beings-with-interests, or what not, and insisting that respecting and promoting their good must be good for us in order for us to have good reason to do it just taints benevolence and justice with an unpalatably narcissistic sort of self-centeredness. To some extent, I think this objection persuades because many of us have encountered people who help others less out of concern for those people than out of concern to prove to themselves and others that they are virtuous, perhaps even superior to others in their virtuousness. I’m happy to agree that if that were the kind of thing Aristotle endorses, he would be urging us to cultivate an unattractive sort of character.

But I suspect that we often overlook the ways in which we benefit directly and non-instrumentally from our own benevolence in part because it comes so naturally to most of us. Excepting sociopaths, most of us often encounter others as people whose well-being is to be respected and promoted quite independently of any instrumental benefits we may gain from doing so. This can easily suggest to us that their welfare has reason-giving force entirely independent from our own. But one way — an Aristotelian way, if not quite Aristotle’s way — to understand our experience of others as to-be-respected is as our attraction to the goods of friendship and benevolence (in their broad senses). To be sure, this is not a good in which the welfare of others is merely instrumental; but it is, I think, a good for the agent. We might find some evidence for this in the extent to which we want and enjoy respectful and benevolent social relations for their own sake. Philosophical literature sometimes suggests that we find these inherently burdensome, because it focuses excessively on cases in which respect or benevolence seems to conflict with some purely self-regarding interest we happen to have. But, to conclude, it seems to me neither implausible nor excessively self-centered to think that rational animals like us benefit in and through respectful and benevolent relations with others and that this benefit supplies us with a powerful reason to cultivate these relations and to resist whatever would threaten them. To complain that this falls short of purely disinterested benevolence is to accept what Rawls, following W.D. Ross, called “the doctrine of the purely conscientious act” (Rawls, A Theory of Justice revised ed., p. 418) — the notion that morally praiseworthy action must be motivated solely by considerations of rightness without any concern for what is beneficial, especially to the agent herself. Rawls rejected this doctrine in favor of the view that we each have a fundamental interest in expressing our natures as rational beings in the form of a commitment to justice, and he is rarely accused of proffering an unappetizingly self-centered philosophy. Aristotle, I think, is no worse off, and perhaps in considerably better shape.

6 thoughts on “Eudaimonism, ‘Egoism,’ and Degrees of Finality in Aristotle’s Ethics

  1. That’s an immensely subtle and clarifying post. Whether or not section III gets Aristotle right, I think it basically nails the central philosophical issue. Many years ago (maybe a decade ago), I heard Terence Irwin defend a very similar sort of view (this was at the Princeton Colloquium on Ancient Phil), but his defense of it was spontaneous, and in response to a question from the audience. It wasn’t as worked out as what you’ve offered. Coincidentally, John Cooper was in the audience that day (as was Allan Gotthelf).

    There’s a lot in your post, and I have a lot to say in response, but before I do, I wanted to get clear on the terminology you’re using and what you take to be the upshot of your view. So, a couple of questions of clarification below. (By ‘clarification’ I mean: how do you, or would you, use the terms, not how do they map onto Aristotle’s claims.) I’m a bit reluctant to pose these questions in this stark way, because doing so almost makes it sound as though I’m administering an exam, but that’s not my intention; I’ve just always thought that discussions about “egoism” were confounded by terminological ambiguities, so that it pays to get clear about terminology from the start.

    1. Is treating someone as “merely instrumental” to one’s good (as you see it) equivalent to treating them as a “mere means” to one’s good?
    2. Is the first grade of finality on Cooper’s analysis of Aristotle equivalent to “mere instrumentality”?
    3. If someone’s good is a “direct source of reasons” for me, can their good at the same time constitute my good? In constituting my good, can their good (causally) promote mine? I ask this pair of questions because if the answer is “yes” in both cases, I no longer see the point of the term “direct” except as a description of the phenomenology of promoting their good. If someone’s good is phenomenologically “direct,” then the need to promote their good is experienced by the would-be benefactor in an undeliberated fashion as something to promote ASAP. But that phenomenological point leaves open the metaphysical/axiological question of whether or not the promotion of their good causally promotes one’s own. Incidentally, in using the term “causally,” I mean to abstract from conceptions of causality. I mean “cause” in any sense of cause, not some narrow sense.

    I personally am agnostic about whether Aristotle was an egoist, and my impression is that sophisticated Objectivist interpreters of Aristotle (e.g., Greg Salmieri, at Boston University) are likewise agnostic. There are (we might say) egoistic elements in Aristotle’s view that make it a precursor of Rand’s, but it’s Rand’s view rather than Aristotle’s that’s full-bloodedly egoistic. Insofar as she expressed a view, I think Rand herself thought that, or something like it.

    Having said that, though, I’m curious what obstacle you see to ascribing egoism to Aristotle on the Cooperized interpretation you offer. Is it simply that the term “egoistic” can be misleading, or brings unwanted baggage with it? I don’t think that’s a very strong objection by itself. To remedy what’s misleading about the term, one can simply define “egoism” narrowly as a technical term (which is what Nagel did, after all, with “altruism” in The Possibility of Altruism). I’m inclined to think that “agent relative welfarism” is potentially just as misleading as “egoism”: supply a sufficiently wrongheaded conception of “welfarism” and you get one problem; supply, say, Scheffler’s conception of “agent relative” and you get another. (Put them together and you get two.) So while there’s not necessarily strong reason to ascribe “egoism” to Aristotle, given your view, there’s no strong reason to avoid the ascription, either. There’s certainly no reason to think that if you ascribe “egoism” to someone, there’s (any longer) any obvious inference to the claim that he’s recommending some obviously disrespectful or immoral treatment of others.


  2. My view owes a great deal to Irwin, though he would reject it, I think; he has recently argued that for Aristotle, acting for the sake of the fine or the noble (the kalon) is an essentially impartial sort of motivation, and he would probably regard my interpretation as too instrumentalist. But since you ask about the terms…

    I’m not sure quite what you have in mind by treating someone as a “mere means,” but if you’re thinking of the sort of thing that the Kantian slogan “treat people always as ends-in-themselves and never as a mere means” wants to exclude, then that’s not quite what I have in mind by treating someone as merely instrumental to one’s own good. Treating someone as a “mere means” in the sense of the Kantian slogan characteristically involves acting with no regard for their good except insofar as it promotes one’s own, much in the way that one treats, say, a coffee maker. In the sense I have in mind, regarding someone else’s good as merely instrumental to one’s own is to be contrasted with valuing that person’s good for its own sake as one of one’s own ends. So purely instrumental regard, on my view, is consistent with attitudes of respect and benevolence, since respecting a person or wishing him well does not require valuing his good as one of one’s own ends; if anything, appreciation for the benefits someone does for me is far more likely to generate and support attitudes of respect and benevolence than the kinds of attitudes often described in Kantian terms as “instrumentalizing.” This is one reason why the concern you gesture toward in your concluding sentences is not among my reasons for being uneasy with the “egoism” label; even if Aristotle, or anyone else, does affirm what I called “reductive, instrumentalist ethical egoism,” holding that we should always treat others as merely instrumental to our own good, it in no way follows that they are committed to endorsing, or unable to condemn, disrespectful or immortal treatment of others, let alone recommending it. I take it, for example, that Epicurus and Hobbes are instrumentalist egoists, and while perhaps neither of them could pass the test set by Gyges’ ring and similar scenarios, they both offer strong reasons for being just and benevolent in many circumstances.

    So yes, “mere instrumentality” here is meant to be equivalent to the first grade of finality in NE 1.7. The trouble, as I see it, with forms of egoism that do not extend beyond instrumental regard for others is not that they must encourage us to treat others manipulatively or cannot support any criticisms of injustice, but that they fail to account for the apparent value and reasonableness of non-instrumental forms of concern and are unable to provide the right kind of account of justice as a virtue. I think Plato had this right; the contractualist account of justice that Glaucon and Adeimantus offer in the Republic is capable of explaining why certain kinds of behavior are just and shows that we have good reasons to practice justice ourselves and encourage others to do so, but it fails to show that we should prefer to be just when we can be sure that we will seem just even while doing injustice. Of course, if reductive instrumentalist egoism is true, then its inability to account for these other notions will not be a flaw. I don’t regard the goodness and reasonableness of justice and non-instrumental regard for others as fixed points with foundational status, but I’m more confident in them than I am in most theories of morality or practical rationality.

    As for your question about the good of others being a “direct source of reasons,” I suppose I see no reason why it would exclude the possibility of the other’s good constituting and/or causally promoting one’s own. It would simply exclude the notion that the other person’s good has reason-giving force only by being somehow connected to one’s own good. But your point against “directness” is more or less the same as mine; when we allow for the ways in which the good of others can constitute and promote our own, and in which our own positive, non-instrumental other-regarding attitudes benefit us, then we do not need to insist that the good of others is a direct source of reasons. We might still retain some notion of directness for the phenomenology, as you say, but we can explain that phenomenology as non-deluded without appealing to an agent-neutral theory of reasons that would make the well-being of others a direct source of reasons.

    As for the point about terminology, you’re of course right that just about anything can be misunderstood. But I think there are two disanalogies between “egoism” and “agent-relative welfarism.” First, most people don’t already think they know what “agent-relative welfarism” is; I try to avoid “egoism” because people tend to suppose that they know what I mean even though they don’t, whereas that isn’t the case with “agent-relative welfarism.” If I tell someone on the street that I am an egoist, she will almost certainly form the wrong impression; if I tell her that I am an agent-relative welfarist, she will likely see that I need to explain that (or, more likely, she will think I am a weirdo and find an excuse not to talk to me). Second, though I think the view that I attribute to Aristotle and find extremely attractive in its own right is aptly characterized as a form of rational egoism and might even fit, at least with the right qualifications, the description of ethical egoism that you’ve elsewhere cited from Gotthelf — “the view that a moral agent should be the ultimate intended beneficiary of every action he takes” — I do not think “egoism” would ever suggest itself as a label for a conception of an individual’s good that gives an important place to non-instrumental concern for others; though certain philosophical articulations of what the term stands for seems to me to be true of Aristotelian (and other varieties of) eudaimonism, I’m inclined to agree with David Norton that eudaimonism is (to quote from Reason Papers again) “a unity from which ‘egoism’ and ‘altruism’ are subsequently extracted and developed as abstractionist fallacies.” So it isn’t just that the term is apt to mislead without lengthy qualification, but that its literal sense seems awkward as a label for a view that emphasizes the value of benefiting others.

    Nonetheless, I readily acknowledge that I haven’t done enough to show that Aristotle does hold this view — though the NE 1.7 passage about finality does, I think, commit him to it — much less that it can address the worries that people like Kraut and Price have. I’ll try to do some of that in the coming weeks, as part of some work that I’m doing on a paper about the scope of justice in Aristotle’s theory, in which I want to defend the view that justice is not limited to the political community but applies in some form in all human relationships. More on that later.


    • Re Irwin on ta kalon: Aristotle’s conception of ta kalon as the motivation for moral action is, to my mind, the major obstacle to describing him as a full-fledged egoist. I like Gotthelf’s conception of egoistic action as action motivated so that the agent is the ultimate intended beneficiary of his or her actions. What that entails, I think, is that to describe someone as an egoist, we have to distinguish two different questions:

      1. What is in fact beneficial for, or good for, the agent?
      2. Motivationally speaking, what is the agent aiming at?

      An action is egoistic if and only the agent aims at what he takes to be beneficial. Regarding the action as beneficial and taking it under that description is necessary and sufficient for the action’s being egoistic. That allows for two possibilities:

      (i) S has some aim X, S takes X-ing to be self-beneficial, but S is wrong about it.
      (ii) S has some aim X, S doesn’t take X to be self-beneficial, but X is self-beneficial.

      On the Gotthelf view, (i) is egoistic but (ii) is not, despite the fact that in (i) the agent doesn’t realize his self-interest, whereas in (ii) she does. There’s some complexity in both (i) and (ii) over what counts as “taking X-ing to be ___” but I’ll gloss over that here. As I see it, the primary obstacle to ascribing egoism to Aristotle is that Aristotle is committed to the claim that virtuous agent acts for ta kalon, but it isn’t clear that when you act on ta kalon, any part of your motivational structure is focused on advantage or benefit. It’s focused on what what is prepon, what is perfectly appropriate to the occasion. I myself don’t see any strong tension between acting self-beneficially and acting on what’s appropriate to the occasion, but it’s not clear to me that Aristotle sees things that way.

      Of course, a rationally egoistic agent ought to aim for (1) and (2) to coincide–to regard as self-beneficial what really is self-beneficial, and pursue it. But my point is, that an action still counts as egoistic if it aims self-beneficially at the wrong thing. I take Rand to have been making this point in the Introduction to The Virtue of Selfishness (p. viii), and I agree with her. The distinction she’s drawing resembles the distinction you’ve drawn between rational and ethical egoism in part I of your post, but it’s not exactly the same.

      I asked the question about “merely instrumental” and “mere means” because Kraut equates the two in one of the passages of his that you cite. The truth is that I find all talk of “instrumental value,” “instrumentalizing,” etc. ambiguous and confusing. One problem is that the discussion revolves around a species of instrumentalization (treating in a merely instrumental fashion), so that there’s the danger of conflating genus and species. I see a subtle ambiguity in your discussion. On Kant:

      Treating someone as a ‘mere means’ in the sense of the Kantian slogan characteristically involves acting with no regard for their good except insofar as it promotes one’s own, much in the way that one treats, say, a coffee maker.

      The problem here is that if we adopt your Cooperized interpretation of Aristotle, and we admit that someone’s being a constituent of your good is compatible with that person’s causally promoting your good, there is no sharp contrast to be drawn between acting with regard for someone’s good qua constituent of one’s own good and treating them as a mere means in the Kantian sense. In both cases, it’s true that you act for their good for your own good. In both cases, your regard for their good is motivated by the contribution that such regard makes to the promotion of your good. There is certainly a contrast to be drawn between a virtuous person’s regard for the good of another person, and and her regard for the good of a coffee maker, but “acting with no regard for the good of others insofar as it promotes one’s own” doesn’t capture it. You can “act with no regard for the good of another person insofar as it promotes your own” by recognizing that as a political animal, their good is a constituent of yours, promotes yours, and gives you reason to promote their good–a reason you wouldn’t have if it didn’t promote your good. Once you admit that non-instrumental regard for another is causally promotive of one’s own good, I think one loses a precise sense of what “instrumental” is supposed to mean. There’s a similar ambiguity in Cooper’s characterization of the first grade of finality:

      a. Goods that are always chosen for the sake of something else and never for their own sake

      We’re apt to think of things like money, but (a) could conceivably be read to apply to the good of others as well. If the good of others doesn’t give us “direct” reason–where “direct” means: bypassing any promotion of one’s own good–then even character friends choose friendship for the sake of something else (the agent’s eudaimonia) and never “for their own sake” independently of any such contribution. Which is not to say that friendship isn’t chosen for its own sake in a different sense, a sense compatible with its contribution to something else (something more final). After all, Nicomachean IX.9 asks why we have friends, and however tortured the chapter is, the question wouldn’t make sense if we insisted that we choose the good of others only for the sake of the good of others, and not because it contributes to one’s own good.

      So I suppose I’m both agreeing and disagreeing with you. I’m agreeing that there are many different ways in which the good of others can constitute and promote our own, but also saying that the instrumental/non-instrumental distinction muddies the waters in clarifying those ways.

      On the term “egoism,” I disagree. There are two potential audiences here–non-philosophers and philosophers. If we’re talking about non-philosophers, “agent relative welfarism” and “egoism” are equally likely to mislead. Non-philosophers hear “egoism,” and hear “predation.” But if they hear “welfarism,” they’re likely to think of “welfare” (in the policy sense: food stamps, etc.) I know you were joking when you said that if you use agent relative welfarism with someone in the street, they’re apt to find an excuse not to talk to you, but I would just take that at face value: that is as much a “wrong impression” as “egoism=predation.”

      If we’re talking philosophers, I think my original points holds. Philosophers equate egoism with predation, but they import their own conceptions of “welfare” into any conception of “welfarism.” In both cases, you have to do some revisionary semantic work.

      I insist on egoism because while I agree that “egoism” elicits wrong impressions, I think philosophy sometimes has to be in the business of eliciting, diagnosing, and correcting wrong impressions of just that sort. People’s reactions to “egoism” reflect a truncated conception of the beneficial relations we can have with others: either we can have extremely reductive, instrumentalized ones–or we must sacrifice our good to others. What is lost is the entire moral universe that falls outside of that false dichotomy. I think Rand was right to want to insist on challenging that in a confrontational way. It’s interesting that in other languages, one doesn’t get the same result as in English. In Urdu, for instance, the word khud means “self,” and “-i” means roughly “-ishness.” In the early twentieth century, the poet-philosopher Muhammad Iqbal articulated a proto-Randian ethics of khudi–literally, of selfishness. But even in ordinary Urdu, khudi doesn’t connote reductive instrumental valuation of others; it connotes integrity. There’s a very contingent history (connected, I think, to Freud) that explains why we 21st century English speakers have come to see egoism as entailing reductive instrumentalism. I think Rand was right to challenge it. Greg Salmieri seems to be suggesting that Aristotle was challenging a different form of it–Aristotle in Socratic mode.


  3. I agree that “acting with no regard for the good of others except insofar as it promotes one’s own” is an unhappy formulation of instrumental regard. Not only does it fail to stand in contrast to what I have been calling non-instrumental regard for others as part of one’s good, but it fails to stand apart from the attitude that I want to call instrumental regard but to distinguish from “instrumentalizing” of the sort that Kantians (rightly) find objectionable. The quasi-Kantian notion (I argue that Aristotle recognizes it and regards it as paradigmatic of injustice, and even connects it with the concept of instruments) picks out an attitude in which I disregard another’s welfare except insofar as it is useful to me in promoting my ends. Insofar as I adopt an “instrumentalizing” attitude towards a person, I do not even see her as subject to any form of benefit or harm that is not connected to her usefulness to me. That is how we see instruments: houses, computers, wine (Aristotle’s example), and so on. We recognize that instruments may be useful to others besides ourselves, and so we do not regard them strictly in terms of their usefulness to us, but we rightly see that for them to be made better or worse is strictly a matter of their being made more or less useful. This is where the instrumental metaphor really belongs, I think, and where Aristotle uses it (indeed, I am not sure it is even a metaphor). But my use of ‘instrumental’ as equivalent to Aristotle’s ‘for the sake of something else and not for its own sake’ follows standard philosophical usage (In my soon-to-be-completed book, I distinguish these and others and argue that Aristotle recognizes a diversity of value and valuational attitudes within what contemporary philosophers typically speak of in terms of ‘instrumental goods’).

    I don’t think we lose any clear and important distinctions between the instrumental and the final by thinking that the virtuous person will regard the good of some others as one of his own ends. To value the good of another as one of my own ends because I benefit in the way I’ve been trying to elucidate is not to value it solely for the sake of something else, because the benefit is not distinct from valuing the other for his own sake. The benefit to me just is the activity in which I relate to the other person; that person and my relation to him will of course benefit me in less direct ways as well, but insofar as I regard my friend’s good as an end, the benefit to me is constituted by my benevolent and caring activity and is not a separate thing that my benevolence causes or promotes. My good consists in my activity, and it is the activity of benevolence — which necessarily involves caring about the person for her own sake — that benefits me.

    This kind of concern and this kind of benefit differ from what we find in “advantage” friendships. When my mechanic fixes my car, he benefits me, but he does not value my good as one of his ends; he values it solely as a means to achieve his own ends, which do not involve me at all. This would be just as true, I think, of an ordinary mechanic and of a rich enthusiast who didn’t charge me a fee but did the work simply because he loves the technical task of fixing cars. When the good for the sake of which he fixes my car is money, this is apparent; he benefits me by fixing my car, but he does so only because it is a means to achieving some good for himself that he wants and that is produced, but in no way even partially constituted, by benefitting me. Unlike money, the good for the sake of which the enthusiast acts necessarily benefits me, but my benefit is not relevant to the description under which he values it. Both of these cases seem pretty clearly distinct from the ones in which someone values another’s good as an end.

    But these sorts of “advantage” relationships are also distinct from the more literal, quasi-Kantian kind of “instrumentalizing.” Valuing someone else’s good and choosing to benefit her solely for the sake of something else is a far cry from regarding her as a mere tool for one’s use. I might benefit my mechanic and care about his welfare only for the sake of the good work he does on my car. But that is not only consistent with recognizing that he has a whole set of interests independent of mine, but gives me good reasons to want him to do well and to take some action myself to promote his interests. This sort of ‘instrumental’ value and regard is not to be disparaged or avoided; it is rather the foundation of social and political life. It would not be sufficient to meet Glauconian sorts of worries about the value and reasonableness of justice, but that is not a good reason to think that it is inherently exploitative or predatory.

    So in principle, I’d be happy to restrict “instrumental” language to the attitude that reduces the good of others to their usefulness and to operate solely with the language of means and ends, perhaps with a distinction between productive and constitutive means. But I think the use of the term to talk about productive means is so standard that I’m not inclined to abandon it altogether. But maybe I should; after all, I think it suffers from the same problems that ‘egoism’ does, and while I’m not completely opposed to the latter, I do prefer alternatives. I’ll have to think about this some more.

    Would you be interested in reading the relevant chapter draft of my book? This exchange shows me just how much I could benefit from your comments.


  4. Pingback: Happy 2015, and some odds and ends | Policy of Truth

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s