Mother Teresa spends her life caring for the poor and the infirm; J. Paul Getty, Jr., spends his life making investments and directing corporations. Although we might be unhappy doing what they do, we assume they are satisfied. Mother Teresa enjoys her work and would be miserable if she had to mastermind corporate takeovers. Getty would be wretched if he had to care for lepers or become a lawn chair salesman.
If we reflect on our own lives and the lives of our friends, we find we are no different. Doing what we want will usually make us happy, whether we want to be benevolent or money-grubbing. Our self-interest is often promoted by intentional action, even action which is presumably morally motivated. Conversely, if we cannot do what we want, we will likely be dissatisfied.
Given these observations, what can we conclude about the relationship between morality and self-interest? Are some acts benevolently motivated, or are they merely disguised forms of self-aggrandizement? The psychological egoists say it is the latter. In their view people always act to promote their own self-interests. This thesis has a surface plausibility and is widely held by non-philosophers. Many philosophers, though, repudiate it. I suppose that is understandable. Some versions are so preposterous that we are tempted merely to scoff at them. Yet far too often ethicists are so anxious to defeat psychological egoism that they fail to notice the germ of truth in it -- a psychological observation with important implications for moral psychology.
In this paper I will identify this insight and argue for its significance to
moral psychology. I will begin by identifying the intuitive appeal of
psychological egoism. I will argue that any univocal, non-trivial statement
of the thesis is false. To do so will require traversing some old ground while
possibly plowing new ground as well. This will enable us not only to specify its
deficiencies, but also to identify and appreciate its strengths.
The intuitive appeal of psychological egoism
Everyone recognizes that she is frequently motivated to promote her self-interest. Moreover, in reflective moments she realizes she is occasionally mistaken when she believes she is not motivated by self-interest. She may initially think she is acting out of concern for another or from commitment to moral principle, yet retrospectively discern that the principal, if not sole motivation, is self-satisfaction, e.g., she expects returned favors, a better image in the community, etc. The egoist thrives on such unmasking of hidden motives. She demonstrates that sometimes an individual ultimately recognizes she was motivated by self-interest, even when she was initially convinced otherwise, and concludes that everyone might be similarly fooled. The egoist thesis is further bolstered by the realization that an individual's intentional actions are rewarding to her in at least some ways, regardless of the motive for them. One who regularly acts morally is satisfied by the realization that she is (and is perceived to be) moral. Loving spouses and parents usually achieve great satisfaction from caring for and supporting others and from being loved and needed. Sacrificial saints reap immense satisfaction from their work. It is difficult to cite a case where an individual acts in ways which will not be at all satisfying to her.
These observations help elucidate the appeal of egoism. Moreover, they must be accounted for by any adequate theory of motivation, either by incorporating and explaining them, or by showing them to be illusory. My contention is that these observations can be best accounted for by the following egoistic-looking thesis: a person will continually engage in an activity only if it has the effect of satisfying what she perceives to be in her own self-interest. Though this thesis is obviously similar to egoism, it differs from it in several notable respects. To understand fully how it differs from egoism as typically described, I must first identify errors in the traditional egoistic thesis.
What's wrong with psychological egoism
Egoism's appeal arises primarily from its ambiguous formulation. The thesis can be explicated in a number of ways. When an opponent objects to one characterization, the advocate shifts to a defense of another. Then, if that version is successfully criticized, still another is forwarded. When the ambiguities are swept away, however, the thesis is demonstrably false. My initial task, then, will be to specify these ambiguities, thereby demonstrating the thesis's deficiencies. The recognition of a single ambiguity will not be sufficient to indicate the deficiencies in the thesis. Nonetheless, when taken in conjunction, they will do so. One quick warning: since it is easy to slide back and forth between diverse formulations, the reader should keep the identified distinctions in mind as she reads succeeding sections.
Motives or consequences
The first ambiguity arises because of the egoist's tendency to blur or ignore the distinction between the motive for an action and the consequences of it. Simply because an act has certain consequences does not imply that the agent was motivated by that consequence. For instance, all acts invariably have consequences of which the actor is ignorant; hence she could not have been motivated to achieve each of them. Moreover, one may do an act which she can reasonably predict will have certain effects, yet still not be motivated to achieve those effects. For example, one might be motivated to cook a special meal for her spouse in order to make him happy, though she would doubtless be satisfied subsequently. What sense could we make of someone for whom it was regularly otherwise: "I am doing something I know my spouse -- the love of my life -- will appreciate, and it makes me absolutely miserable.") To claim she cooked the meal in order to gain the satisfaction would be to misdescribe the case. That is not to say that some cases cannot be so described. We all know people who purportedly act for others when, in fact, they are only concerned about themselves. Just because some people are so motivated, however, we should not conclude that all are. There are clearly differences between these cases, and any adequate theory of motivation should acknowledge them. We all can discriminate between those who are motivated to help others yet are consequently satisfied by doing so, and those who aid others merely as a means to their own self-interest. The difference between these people can be explicated not only in terms of their motives but also behaviorally. The first person would have cooked the meal even if, in a certain case, she had known that it would not directly satisfy her; the second wouldn't have. Put differently, the first individual is likely to assist others in a variety of circumstances. The second would aid others only when it is in her identifiable self-interest to do so.
One self's interest or one's self-interest
These observations illuminate an important related distinction the egoist obscures: one between an action's being in one's interest and its being in one's self-interest. This distinction further elucidates the difference between the above-described individuals.
The first person has an interest in her spouse. She wants to make him feel cherished: she wants him to be happy. Likely she will be pleased if she helps him satisfy his desires. But that does not suggest that she was motivated to please herself. The second individual wishes to promote her own interest; she "helps" others only as a means to that end.
Psychological egoists gloss over this important distinction. They begin with the evident fact that individuals pursue their interests. Whenever an individual can articulate a reason for her actions they are, ipso facto, her reasons. To say this, however, simply identifies the locus of the interests. It says nothing about the nature or object of them. Some of a person's interests may be self-interests, interests merely in her own welfare; other interests, though, may be interests in other people. Individuals are usually self-satisfied when they achieve their interests, including their interests in other persons. But that does not imply that the motivating interest was self-interest. Again, the difference is describable both in terms of present motives as well as behaviorally. An individual seeking her self-interests would gladly forgo her efforts to help another if she could be satisfied in some alternate way; the one motivated to help the other would not. Consequently, if psychological egoism is understood in the first way, as saying people always act to promote their interests, it is true, but uninteresting. it simply states that anyone's interests are hers -- that person's and not someone else's. It does not tell us what interests she has. In particular, it does not tell us if some of her interests are interests in other people. If egoism is understood in the second stronger way, evidence strongly suggests it is false, since individuals do have interests in persons besides themselves.
These observations help us to articulate a fundamental weakness of egoism as a theory of human motivation. Egoism treats pleasure or happiness as a uni-dimensional state, a determinate end which people seek like moths seek the flame. They wander through the world in search of happiness and do anything which will serve as a means to that all-encompassing end. But happiness is not a state or thing which one automatically achieves by following a certain recipe. Rather, happiness is primarily a byproduct of activity, not a thing to be directly pursued. Of course, we may wonder what makes some actions happiness-producing. A complete answer to that question is beyond the scope of this paper. To obtain a partial answer, however, we must only realize that happiness can come in all shapes and flavors. Humans are malleable creatures, quite capable of finding. happiness by doing any of a wide range of activities.
Of course there are natural limitations on what activities can be satisfying. I suppose they are minor, however. Personal, historical, sociological, and psychological evidence suggests people pursue activities for a number of diverse reasons, and are capable of finding a plurality of them satisfying. Although the specific reasons for acting may vary widely, those that bring satisfaction seem to have a common denominator: if people do things they desire to do, they will be content. As Derek Parfit puts it, "what is it our interests partly depends on what our motives and desires are."
Put differently, people have strikingly divergent interests. Some are self-interests; others arc nonself-interests. Fulfilling these latter interests will usually be satisfying, but they will be so only because they are her interests; they are not her interests (let alone her self-interests) because they are satisfying. Since any number of actions can be rewarding if a person desires them, the general explanation for pursuing one particular action instead of another cannot be: because it is satisfying. Admittedly we need to ascertain why people have the interests they do. That is evident, though, is that the egoist cannot provide that explanation. She can point out that both Gandhi and Idi Amin acted in accordance with their interests, but she cannot explain why each had the interests he had. Nor could she have helped either to know which interests he should have developed.
Real interests or perceived interests
Thirdly, egoism is ambiguous between the claim that individuals always act (whether they realize it or not) to promote what is in their interests and the claim that they act to promote what they take to be in their interests. The egoist moves back and forth adroitly between these alternatives, making the; thesis appear explanatorily potent. However, neither alternative is singularly plausible. The first alternative is faulty on empirical grounds. People often fail miserably to promote their own interests, even when others would agree they are trying to promote them. They may obtain what they want, yet remain unsatisfied. It cannot be the case that people always aim to, promote what is in their interests.
The second alternative is likewise implausible, particularly once we utilize the aforementioned distinction between something's being in one's interest and its being in one's self-interest. Although it might be true that an individual always acts to promote what she believes to be in her interest, it is not the case that she always acts to promote what she believes to be in her self-interest. Not infrequently an individual believes herself to be acting from motives other than promotion of self-interest. Even if she is sometimes mistaken, that would not bolster the egoist's case.
Successful disclosure of hidden motives would show only that an individual was mistaken about what her motive really was. It would not show she was mistaken about what she believed her motive to be. Since she believed she was motivated to promote her non-self interests, the second alternative is likewise defective.
Always self-interest or only self-interest
Finally, the thesis is ambiguous between the claim that people always acts to promote their self-interest to some degree or another, and the claim that they always choose that action which will promote their self-interest better than any known available option.
On the first interpretation the thesis might turn out to be true, though even if it were, the thesis would not be egoistic as traditionally understood. Everyone recognizes that promotion of self-interest is often a motive for acting. Even when a person is motivated to help others she may, at the same time, also be motivated to promote her self-interest. But why should this be thought to support egoism? Egoism is implausible once we realize humans operate from multiple motives--even if one of them is self-interest. In fact, if we admit there are any motives besides self-interest, egoism, in the strict sense is undermined. Consider, for example, cases where agent A has two options: option X promotes her self-interest, but option Y does so even better and she realizes this. If self-interest were her only motive she would automatically choose Y. However, is she has any non-self interests, then at least in some cases, the slight increase in self-interest promoted by Y would be outweighed by a combination of non-self interests promoted by X. In such cases self-interest would not be the decisive motivating factor. Egoism must thus be understood as making the stronger claim: that people always choose that action they think will promote their self-interest better than any known available option. But this stronger claim appears to be counter intuitive, particularly when coupled with previous objections to egoism. For even when people report t hemselves as pursuing their self-interest, many would claim to be likewise pursuing non-self interests. And, again, though they might be mistaken about their real motive, they are not mistaken about what they take their motive to be. Hence, they are not acting in ways they believe will promote their self-interest better than any known available option.
Consequently, the precise statement of a non-trivial form of the thesis is this: all human beings always act to promote only what they believe to be in their highest self-interest. When stated in this precise way, the case against psychological \egoism is powerful. Yet belief in the thesis persists. Why? Is it because people are irrational or uninformed? Doubtless some critics will think so. I would prefer to think, however, that people retain allegiance to this thesis because it appears to embody the insights isolated at the beginning of this paper. It is high time to elaborate on these insights and state a rough theory which incorporates and explains them. I will not pretend to provide an indubitable argument not a final statement of the revised thesis; that would require a complete psychology. Nonetheless, I will offer arguments indicating its plausibility.
The Insight of the Thesis
The embedded insight in psychological egoism is this: a person will continually engage in an activity only if it has the effect of satisfying what she perceives to be in her self-interest.
This thesis differs from the traditional version of psychological egoism in three significant respects, all of which were isolated in the previous discussion. First, it does not assert that a person is always motivated by promotion of her self interest; it only claims she will not continually act in ways that do not have the effect of promoting her self-interest. That is, even though self-interest may not motivate an individual, the achievement of those interests will psychologically reinforce that behavior. As Daniel Dennett puts it, 'the Law of Effect (...that actions followed by reward are repeated...) is not just part of a possible explanation of human behavior, but of any possible adequate explanation of behavior.' (8)
Second, the revised thesis recognizes that some of a person's interests are nonself-interests. That is, an individual may have interests in other persons besides herself; nonetheless, when she realizes those interests, her self-interest is usually advanced as well.
Third, it does not require that self-interest be the decisive, most weighty interest. Her nonself-interests will at least sometimes predominate.
This thesis is plausible when seen against the backdrop of my previous arguments against psychological egoism. moreover, it accords well with the principles of human motivation cited at the beginning of the paper. t recognizes that humans are often motivated by self-interest, and that even when they aren't, they are customarily satisfied by realizing their nonself-interests.
Finally, it is consistent with all current psychological accounts of human behavior and evolution. It seems apparent that we could not have survived as a species without this psychological predisposition. In fact, once clearly stated, the thesis is evident. Perhaps many philosophers recognize the truth of this claim, would even find it trivially true. Nonetheless, ethicists and non-philosophers alike seem to ignore it when addressing moral questions. It is regrettable that they (we) do. The thesis has obvious and direct implications for moral psychology and moral theory.
It is immediately apparent, for example, that on this conception an individual will not continue to act morally unless doing so eventually has the effect of promoting her self-interests, even if it does not promote them better than available alternatives.
That is not to say that an individual is always motivated by self-interest or that satisfaction of self-interest is a consequence of every moral action. Certainly not. It only asserts that we are psychologically constituted to decrease an activity, including moral activity, if we are not ultimately satisfied by it. Consequently, moral psychology should not be primarily concerned with telling us how to battle our natural inclinations -- though, as I shall argue, that is something we must sometimes do. Rather it should tell us how to mould our interests so that they coincide with the demands of morality. It should tell us how to make ourselves and others develop nonself-interests. If such training succeeds, everyone's ability to satisfy these nonself-interests in morality will likely have the effect of satisfying her self-interests. Regrettably that is an unattainable ideal. Some people will never find the moral life satisfying. That is why we need laws to enforce a minimal level of decency. Legal sanctions are designed to make compliance attractive.
Most individuals in a well-ordered society, however, are not motivated primarily by fear of punishment -- or if they are the society will not long survive. Most citizens will comply voluntarily with community moral standards. There are several psychological observations which explain why -- observations which will further elucidate the modified egoistic thesis.
I would conjecture that for most human beings in normal circumstances, being moral would be part of the most satisfying life they would lead. However, this conjecture need not be true for people to act morally; it need only be true that living a morality can be generally satisfying. Considering and looking after the interests of others often does bring (to borrow Nowell-Smith's phrase) a certain glow unattainable in other ways. Opening oneself to love and caring allows one to enjoy human satisfactions attainable only in concert with others. If course we must again note the quasi-paradox noted above: that if one tries to be loving only to gain that satisfaction, the attempt will probably backfire.
The individual who feigns interest in others as a means of promoting her self-interest will fail. She must expand her moral horizons and cultivate interests in others. Then she may find she is satisfied by realizing these nonself-interests.
The best way, therefore, to get people to act morally is to encourage or train them to desire to act morally -- to inculcate an interest in being moral. If, one succeeds, there will be a motivation to pursue moral actions (say, being sensitive to one's friends, or more understanding of individual differences) even when one is temperamentally disinclined to do so. An individual so trained will be willing to pursue a certain moral action simply because it is moral. Moreover, she will be satisfied when she realizes her interest in being moral, even if the action does not directly promote her self-interest. This "piggy back" satisfaction, however, will likely be eventually in, self-sufficient unless the activity becomes more directly satisfying.
This is where habit enters. For as Dewey says, "Man is a creature of habit, not of reason nor yet of instinct." Habits are established by the repetition of an activity. Moreover, they usually become satisfying; it matters not whether the habit is inculcated by parental training or personal will. Furthermore, since human beings are quite malleable, they are capable of numerous different habits, and, given the proper training, finding most of them satisfying. Witness the diversity of manner within our culture, the disparity of lifestyles between cultures. People living radically different ways have managed to attain personal satisfaction. How? As Dewey explains, "Almost any activity continued until it becomes a habit becomes pleasurable and when discontinued becomes a source of pain and dissatisfaction." This does not imply that all lifestyles are equally satisfying. It does suggest, however, that within a broad range people can choose any number of options, yet still attain satisfaction from them.
The egoist misses this important point. She cannot explain why one selects one of these lifestyles over others. Since many of these options could be satisfying to the agent if she adopted them, then unless she chooses the option she thinks most satisfying, it cannot be that she chooses the option she does in order to be satisfied. Other motives must be operative. Although I would not venture to give a complete explanation of such choices, I would say that some people make them on moral grounds; that is, they develop interests they think they morally should have. After doing so they can become habituated to the moral behavior, and the combination of the power of tidbit and the intrinsic rewards of being moral increases the probability that she will continue to act morally.
On the other hand, merely imploring an individual without the requisite training to be moral will likely produce nothing more than a frustrated individual who makes vain attempts to alter her lifestyle. To assume otherwise "is the greatest bar to intelligent social progress. It bars the way because it makes us neglect intelligent inquiry to discover the means which will produce a desired result, and intelligent intervention to procure the means.
In short, it leaves out the importance of intelligently controlled habit." Yet I fear that is what most philosophers and (moral teachers) do. She can encourage individuals to be moral, yet have offered no guidance on how to do that. Moreover, we have so denigrated self-interest that when an individual is satisfied by acting morally, she assumes her behavior is morally suspect. It is imperative to note that when moral training succeeds, the line between one's interests and self-interest blurs. When an individual makes another's interest her own, then his satisfaction will affect hers. This is especially true among intimates and family members. His obtaining what he wants may be just as or more important to her own happiness as is her achieving what she wants for herself. Their interests may become so pervasively intertwined that it is difficult for her to distinguish his welfare from her own, and vice versa.
The egoist notes this common phenomenon, but focuses inordinately on the fact that one's self-interest is sometimes promoted by forwarding the interests of another. In particular, the egoist glosses over the fact that her self-interest is satisfied in this way only because she first developed a nonself-interest in her intimate. As long as we are attuned to the genesis of overlapping interests, we can recognize the superiority of the modified thesis as an explanation for this phenomenon.
Let me briefly describe a case that illustrates the thesis. Suppose someone who has been reared a carnivore confronts the familiar arguments for vegetarianism. After some mental wrestling the person decides -- rightly or wrongly -- that she ought to be come a vegetarian, even though she rather enjoys the taste and texture of meat. If the modified thesis is correct, this person will not make an immediate conversion; she will not turn into a completely committed vegetarian over night. The carnivorous habit is too strong; breaking it is too painful. Moreover, even if this person manages to live up to her convictions temporarily, she will eventually fail unless she comes to be directly satisfied by vegetarianism. As it turns out, the moral desire to be a vegetarian, and the inculcation of the new habit (if the practice is sustained), will likely enable her new eating habits to become satisfying. If, for whatever reason, she does not, she will eventually give it up. Either she will continue to think she ought to be a vegetarian and feel guilty for being weak, or she will manage to "convince" herself that her original assessment of the arguments was all wrong -- that morality does not really demand that she be a vegetarian.
We can further illuminate the interplay of habit and duty in the moral life by briefly examining the views of two moral philosophers typically seen as polar opposites: Aristotle and Kant. Both are trying to isolate important features of the moral life, both overemphasize one aspect of it. Consider first the role habit plays in the Aristotelian conception of the moral life." For the things which we have to learn before we do them we learn by doing: men become builders by building houses, and harpists by playing the harp. Similarly, we become just by the practice of just actions, self-controlled by exercising self-control, and courageous by performing acts of courage. In short, the good person for Aristotle is the one who does good acts habitually.
Kant, on the other hand, denigrates the person who acts habitually even if the acts are normally thought to be good. Goodness is found only in the good will. Acts done habitually are not good; only those done out of a sense of duty are.
Despite the obvious differences between these philosophers' views, there is a way in which they can be harmonized. Consider again our story of the carnivore-turned-vegetarian. At some point in this person's life, she has a conflict between duty and inclination. Thinking that duty demands a change in eating habits, she gives up meat. She is acting out of duty, not inclination.
Here Kant has isolated an important feature of the moral life. Morality does occasionally conflict with inclination and self-interest. If our friend were to simply act habitually she would not be doing what she thinks she ought.
Once she has begun to make the change, however, the situation differs. Within a few months our vegetarian becomes acclimated to her new diet. It now seems normal to her; the thought of eating meat has steadily become less appealing. Soon her inclination is to be a vegetarian. She now does habitually what she once forced herself to do, and she is satisfied by doing so. If my revised egoistic thesis is correct, she would not have continued to do so unless the diet had become satisfying -- in this case by her becoming habituated to the new eating style.
Hence the insight from Aristotle. Since we are largely creatures of habit, it could not have been any other way. To the extent that Kant misses this important fact, he has gone awry. But if he has gone awry it is not because he thinks morality demands that we sometimes act against inclination -- for surely we must. Rather it is because he fails to recognize that we are psychologically so constituted that if we become habituated to an action, particularly if it is one we desire to perform (say, in order to be moral), then we will find that action satisfactory and do it out of inclination.
The moral life, then, is properly seen as a cooperative venture between duty and habit. We are, from youth, taught certain moral principles. If our training is successful, then we will habitually act in ways we were trained to act. Would it be weird and regrettable if it were otherwise? Imagine how awful it would be if, when meeting a close acquaintance on the street, we had, in each and every case, to make a conscious decision to be friendly.) If we were taught the proper moral rules and attitudes, so much the better for us and everyone else. Nonetheless, it is just where habit fails, e.g., because we were improperly trained, that duty need step in. He must, at least on occasion, fight inclination out of a sense of duty. Duty should provide the impetus to inculcate a new habit, one which, we hope, will enable us to be more moral.
Both sides of this tandem are essential. Without either we would be morally frail. And without understanding the germ of truth in psychological egoism we would fail to perceive and appreciate the proper interaction of duty and habit in the making of the moral life.
 Copyright @1988 by Hugh LaFollette.
 In the future I will simply refer to "psychological egoism" as "egoism" to avoid cumbersome phrases. The reader should be careful not to assume that I am talking about ethical egoism.
 I am indebted to Max Hocutt for this way of making this distinction. See his fine paper, "Interests, Self-interests, and Egoism (unpublished).
 We can glimpse here the confusion because or our normal way of speaking about means and ends. As Dewey properly notes in Human Nature and Conduct (New York: Modern Library, 1922), pp. 207ff. human ends are not identifiable concrete points we pursue by any available means. Ends are only foci for human activity. An end is an abstraction from a series or means (itself an abstraction). For instance, if I say I want to reach some end, it is normally identify the end and then search for plausible means to secure it. Indeed, if I seek the end, I seek the means as well. For instance, if I seek to be an attorney, I desire, as part of that end, the means to achieving it, namely, attending law school. To say otherwise -- to claim that being an attorney is an end of mine, yet I don't wish to go to law school -- is simply an illusion.
 "We all have an immense capacity. . . for modifying our behavior in response to what we learn. The result is a remarkable behavioral flexibility.... We are able to devise alternative patterns of 'behavior' for virtually every circumstance or lire...." Gerhard and Fran Lenski, Human Societies (New York: McGraw Hill, 1978). Also see Daniel Dennett's discussion of the evolutionary basis or human adaptability in chapter three of Content and Consciousness (London: Routledge & Kegan Paul, 1969).
 Reasons and Persons (Oxford: Clarendon Press, 1984), p. 87.
 "If I want to relieve the stress or the beggar, I want to relieve the stress of the beggar; I do not want my own happiness, pleasure or satisfaction. Likewise, if I like giving pleasure to my sick aunt, what I like is giving pleasure to my sick aunt, not my own liking, or even the glow I might receive from being benevolent. Indeed, if the latter were what I really liked, I should be doomed to eternal disappointment, because I could never get just that glow unless I acted for the sake of giving pleasure and not for the sake of the glow." Patrick Nowell-Smith, Ethics (Baltimore, MD: Penguin Books, 1954), p. 142. This so-called paradox of happiness is also nicely summed up by Nathaniel Hawthorne: "Happiness in this world, when it comes, comes incidentally. Make it the object of pursuit and it leads us on a wild-goose chase and is never attained. Follow some other object and very possibly we may fond we have caught happiness without dreaming of it."
 I have already argued that, in its strictest formulation the egoist should be talking about promotion of one's self-interest. But I shall continue to play loose with this term, as does the egoist, until the completion of this particular argument.
 Brainstorms (Montgomery, UT: Bradford Books, 1978), pp. 72-3. See also Michael Slote, "An Empirical Basis for Psychological Egoism," Journal of Philosophy, vol. 61, no. 18 (1964).
 See Derek Parfit's discussion in Reasons and Persons of the role of coercion in solving many-person's dilemmas.
 Hume's account of the "circumstances of justice"' is applicable here. For a more recent discussion, see John Rawls, "The Circumstances of Justice," in A Theory of Justice (Cambridge, MA: Harvard University Press, 1971).
 Human Nature and Conduct, p. 18.
 Lectures on Psychological and Political Ethics: 1898 (New York: Hafner Press, 1976), p. 23.
 See note 3.
 Human Nature and Conduct, p. 28.
 Nicomachean Ethics, Book 11.
 Bishop Butler may well have realized this. Although Butler is known for his rejection of psychological egoism, he also says: "Let it be allowed, though virtue or moral rectitude does indeed consist in affection to and pursuit of what is right and good as such; yet, that when we sit down in a cool hour, we can neither justify to ourselves this or any other pursuit, till we are convinced that il will be for our happiness, or al least not contrary to it." Butler's Works, ed. W. E. Gladstone (Oxford: Oxford University Press), 2: 206.
 I would like to thank Jeffrey Gold, George Graham, Gordon Graham, James Rachels, Robert Simon, and especially Joan Callahan and Eva LaFollette, for helpful comments and criticisms on earlier drafts of this paper.