What's Wrong With Egoism?

[Summary: I attempt to respond to Bill Dwyer’s request that I identify what I think is wrong with egoism. I begin with remarks on what I do not think is involved in the rejection of egoism, follow with some reflection upon just what is being requested, sketch what I take to be involved in egoism, and then proceed to outline some of the reasons (I don’t think there’s just one) that egoism should be rejected. What follows is long. I have inserted section-headings to break it up into more manageable units. This is lightly edited from something that was presented to an electronic mailing-list.]



What’s Not Involved In Rejecting Egoism

Given some of my earlier remarks, Bill Dwyer raises the following very reasonable question: “I am still unclear as to what he thinks [that is, what I think] is fundamentally mistaken about a doctrine which holds that a person’s ultimate value is his own happiness? In other words, what does he think is the essential error in an ethics which rejects self-sacrifice as immoral?�

I’d like to begin with some ground-clearing, specifically, with identifying two things that are not meant (at least by me) by the rejection of egoism. First, I do not mean to be saying that self-interest is a bad thing or unimportant or that it has no proper role in a good life.

As a first approximation, we can say that egoism is the theory that all that one ought to do derives directly or indirectly from one’s own interests. Therefore, denying that theory doesn’t necessarily mean that nothing one ought to do derives from self-interest, but rather that self-interest isn’t the only thing that counts (ultimately). To quote a famous non-egoist:

Certainly, our well-being and woe count for a very great deal in the appraisal of our practical reason and, as far as our nature as sensible beings is concerned, all that counts is our happiness if this is appraised, as reason especially requires, not in terms of transitory feeling but of the influence this contingency has on our whole existence and our satisfaction with it; but happiness is not the only thing that counts. The human being is a being with needs, insofar as he belongs to the sensible world, and to this extent his reason certainly has a commission from the side of his sensibility which it cannot refuse, to attend to its interest and to form practical maxims with a view to happiness in this life and, where possible, in a future life as well.

That’s Immanuel Kant from the 2nd critique (61, Prussian Academy Edition, emphases in the original).

Let me emphasize this further, particularly in light of Kurt Keefner’s recent statement that “Altruism is not the only logically possible alternative to egoism, but it is currently the most prominent alternative.� Altruism, as Rand defined it, is not the only alternative to egoism. Neither is it a prominent alternative – not now nor has it ever been. Rand’s criticisms of that position may be entirely decisive, but they are of negligible importance, because what she called altruism is a view that almost nobody has ever held. I cannot think of a single contemporary moral theorist who holds it nor is there any important philosopher in history who held it. It is a view that exists in the imagination of Objectivists, not in the real world.

Second, saying that something counts morally besides self-interest is not to say that people should be forced to contribute to the good of others (or to anything else). The question what you ought to do is distinct from the question what you should legally (or otherwise) be forced to do. I mention this point because, in my experience, many Objectivists overlook it. More precisely, they see the point in one connection, but fail to see it in connection with the question whether we have any obligations to others not derived from self-interest. That is, they realize that saying that people ought to pursue their own interests does not imply that those who do not may be forced to. They do not favor, for example, laws banning over-eating, despite compelling evidence that over-eating is not in one’s interests. (For the record, more people die of health problems related to over-eating than of smoking, drinking and illegal drug use combined.) However, when someone asserts that we have some obligation to others, they are quick to infer that the person who says it is anxious to force people to fulfill that obligation. In fact, the alleged implication follows no more in one case than in the other. Accordingly, when I speak about obligations to others that are not derived from self-interest, I am speaking about what people morally ought to do, not about what they ought to be forced to do.

Is Anything “the� Essential Error in Egoism?

With those remarks as background, let me return to the question Bill asked. He asks me to identify egoism’s “essential error.�

I doubt the question has an answer in this form. At least, on what seems to me the most natural interpretation, I don’t see that there’s an answer. Somewhat differing positions (even on this list) have been characterized as versions of or even the only version of egoism. In addition, different egoists hold the position for different reasons. I don’t expect to find some one thing that is “the� essential error in egoism. I do not even think that all egoists are mistaken in any sense that implies that they have made a mistake in adopting egoism. A person may reasonably adopt a false theory if, for example, it seems better than alternatives known to her and if she doesn’t know of any counter-arguments that settle the matter against the theory in question.

What Is Egoism?

Now, since I have alluded to differences among egoists as to what egoism is, let me try to offer a characterization of it that I hope will prescind from most of these differences. Bill refers to it as “a doctrine which holds that a person’s ultimate value is his own happiness.� Others here would say that it holds that one’s own life is the ultimate value. Still others would agree with Bill’s proposal but construe “happiness� hedonistically as referring to an optimal balance of pleasure over pain (I don’t know whether Bill takes it this way or not). I will try to avoid these debates by saying that egoism holds that it is one’s own interests that are of ultimate value, leaving it to different egoists to provide their own accounts of what is in one’s interests.

This needs further explanation in two ways. One is that “ultimate value� has to be construed as agent-relative – as referring to ultimate value to or for the agent. Otherwise, the egoist is caught in the contradiction identified by Moore: He will be saying that one thing, his own interests, is the only thing of ultimate value, and also that it is not the only thing of ultimate value, since he does not believe that other people should put his interests ahead of their own. The other is that the interests said to be of ultimate value to the egoist must not be moralized. I explain what that comes to and why the egoist needs a narrow account of interests in Who is an Egoist? What are Interests?

Second, I take egoism to be a theory that is both mandatory and universal. By calling it mandatory, I mean that it does not merely say that a person is (morally) permitted to take his own interests as having ultimate value or to pursue them (and perhaps permitted not to do so); it holds that he is rationally or morally required to do so. By calling it universal, I mean that this requirement is addressed to every person, not just to some: it is not the case that there are some moral agents who, for some reason, are beyond the scope of the requirement.

Within this framework, we can distinguish two different kinds of egoist positions, which I shall call Narrow and Broad:

Narrow Egoism holds that all permissible actions, plans of action, principles of action and projects must be derived from considerations of self-interest. The ultimate justification of any of these must be that it serves one’s interests to take that action, adopt and execute that plan, accept and live by that principle, engage oneself in that project. I believe this is the position that Rand held:

The Objectivist ethics holds that the actor must always be the beneficiary of his action .... (VOS, p. ix, emphasis added)

and

[Rationality] means a commitment to the principle that all of one’s convictions, values, goals, desires and actions must be based on, derived from, chosen and validated by a process of thought .... (VOS, p. 26)

But this is not the only position that can reasonably be said to hold that one’s own interests are of ultimate value. Instead, one could accept Broad Egoism which does not insist that all actions, plans, principles, etc., be justified by their promotion of one’s interests (though a Broad Egoist may expect that they typically will be). According to the Broad Egoist, however, actions, plans and so on may be adopted without explicit self-interested justification provided that they are not in conflict with one’s interests. One’s interests can act as a veto on any other considerations. Thus, a Broad Egoist might say that he does not need to show that parenting serves his interests, just that it is not inconsistent with his interests.

More briefly, a Narrow Egoist says that one must always act for one’s interests; a Broad Egoist says that one must never act against one’s interests.

On either a Narrow or a Broad view, there will have to be some conception of what is in one’s interests. I shall call this a list. There may only be a single ultimate value on the list (such as life or happiness) or there may be several, such as longevity, health, happiness and perhaps a few more, together with some function or ordering principle for combining them or arbitrating in case of conflict among them. In either case, if egoism is to be an interesting theory, the list needs to have at least two properties. First, the list must be relatively short. (Obviously, a one-item list qualifies.) The interest of egoism as a moral theory is, in part, its claim to be able to derive or give support to other moral claims on the basis that accepting and acting on those claims serves the agent’s interests. The longer the list becomes, the less interesting that kind of support is. Second, it must be initially plausible (or else derivable from something else) that the items on the list represent things that are in one’s self-interest. Without that constraint, we could just put any item we wished on the list – respect for rights, concern for the general welfare, compassion for the needy – and call the resulting theory a version of egoism.

The First Problem: Lack of Support

In several places – on this list, in the paper on Rand’s ethics that I have referred people to, and in others less readily accessible – I have claimed that specific arguments in favor of egoism fail. I will not here try to reproduce those arguments, but I think the conclusion I reached can be generalized. So far as I know, no one has ever presented a sound or compelling argument in favor of egoism. Every attempt that I have seen falls into errors of one sort or another. (For a discussion of some common errors, see Happiness and Ultimate Goals.)

This does not show that egoism is mistaken. It might be that there are no sound or compelling arguments for any other ethical positions either.

But, it should be recognized that the absence of such arguments for egoist or other ethical positions does not mean that there is no way that a moral position can be supported. In ethics, as elsewhere, we have to start where we are. And “where we are� includes our possession of moral beliefs that we would not easily abandon and our engagement in practices that presuppose the correctness of moral beliefs. There still might be an argument in favor of some moral position that proceeds from our considered moral beliefs combined with other things we know to the conclusion that a given theory best harmonizes those prior moral beliefs with one another and with other relevant knowledge. Such a procedure is not necessarily bound to what we already happen to believe, because there may be mutual modification of prior beliefs in light of the best theory and of the best theory in light of prior beliefs. Prior beliefs constrain what an acceptable theory will look like, and the best theory we can construct may provide a reason for rejecting or revising prior beliefs.

I don’t know of any impressive argument of this sort in favor of egoism, either. But even if I am correct in claiming that there is not, that still doesn’t mean that egoism is incorrect.Presumably, if it is correct, it would be better supported than any other theory if we had access to and could properly evaluate all possible arguments in favor of different theories rather than just the ones that have actually been presented.

My present claim is somewhat more modest than any direct inference that egoism is mistaken. It is that, given the present state of the discussion, there is no presumption in favor of the correctness of egoism.

Let me try to spell out what I mean by that a bit further. If there were a sound argument for egoism, that would, of course, settle the matter in its favor. If there were a sound argument for some other moral theory that would settle the matter against egoism. But (I claim) there is no sound argument for egoism, and I am willing to assume for the sake of argument that there is no sound argument for any other moral theory either. (In passing, note that these claims (a) are not equivalent to saying that there are no sound arguments against one or more of these theories. One might be able to rule out some theories as unacceptable even if one didn’t have a direct argument for some particular alternative. And (b) there might be reasonable support for a position that falls short of being conclusive.) If this is so, there is a sense in which the available options in moral theory are competing on a par: Each is an attempt to provide the best understanding of the moral beliefs and convictions that we already hold or that, it can be argued, we should hold in order to have a coherent overall theory. But there is no presumption that any adequate theory must be egoistic. If there are moral convictions that we, after critical reflection, are still disposed to accept but which appear incompatible with egoism or difficult to reconcile with it, there is no reason to assume that they must somehow be reconcilable with it or else should be rejected. Egoism is no more a favored starting point than any other moral theory. It may be that the best explanation for our having and being reluctant to relinquish such convictions is that egoism is an incorrect theory. Those convictions may be responsive to morally salient facts that egoism does not properly handle.


The Second Problem: Conflict with Considered Moral Convictions

So far, this has been quite abstract. Are there any moral convictions, difficult to dismiss as errors, that are also difficult to reconcile with egoism (preferably, for the sake of argument, convictions that egoists find themselves accepting)? I think there are. I will begin by mentioning four cases. In discussing these cases, the convictions to which I will appeal are, I think, convictions that nearly all egoists share with nearly all non-egoists. I will mention, but not discuss here, that there are special problems for egoists who are also Objectivists in that their commitment to egoism (I would argue) is not actually consistent with the theory of rights that they also endorse. If they wish to keep egoism, they should give up the rights theory; if they wish to keep the rights theory, they should give up egoism. (I think they should do the latter.) Some discussion of this point may be found in my paper on Rand’s ethics and, at greater length, in Egoism versus Rights, in the Spring 2006 issue of the Journal of Ayn Rand Studies.

The First Case: Do Others’ Interests Matter in Their Own Right?

One example is provided by the Kitty Genovese case that Chris Cathcart has mentioned on this list. Here is a slightly edited version of what I said elsewhere (followed by some additional commentary):

On March 13, 1964, in Queens, New York, Kitty Genovese was raped and murdered while at least 38 people in a nearby apartment building witnessed all or part of the attack. While she was being attacked, she was screaming for help.Yet no one even took the trouble to call the police until 35 minutes after the incident began. By that time, it was too late for Kitty Genovese. The ethical egoist doesn’t have to say that none of those people did anything wrong. But it looks like he does have to say that if any of them were doing anything wrong, it was because they were not doing as well as they should for their own interests. Maybe, somebody could have gained a reputation as a hero, gotten his picture on the front page of the newspaper, by intervening. Maybe, someone was just annoyed by the screaming and could have stopped it by calling the police earlier. Calling the police might have been a better solution than turning up the volume on the television or shutting the window. Whatever reason those watchers might have had for intervening – if the ethical egoist is right – it would not have had anything to do with Kitty Genovese’s interests or welfare or life. Those just drop out of the picture as unimportant.

I think most people would say both that someone should have taken the trouble to call the police and that the reason for doing so was that Kitty Genovese’s interests were at stake, not that their own might have been advanced by making the call. Her life, her interests, her welfare count in their own right, not just because of the way they happen to mesh with the interests of others.

Note that saying that her interests count is not saying how much they count, or that people would be obligated ever afterward to make sure she lived well, or even that people were obligated to make large sacrifices then and there. Placing a phone call is hardly a large sacrifice. It is just saying that the people who heard her screams had some reason to act that was based on Kitty Genovese’s interests rather than their own.

Since this is not a case in which the interests of some are sacrificed for others, a Broad Egoist could allow that at least some of the people who heard her might have had a reason in terms of their values, goals, etc. to place the call. The problem with that is that it seems to treat Kitty Genovese’s claims upon those who heard her screaming too casually. It says that a person is morally permitted to take her interests into account, but does not have a way of saying that he should have done so. Putting this the other way around, if we imagine that an egoist was watching who didn’t happen to have adopted some value or project that committed him to taking even minimal steps (like calling the police) to come to her aid, but instead found the scene entertaining – then he would have been doing nothing wrong (or might even have been doing something wrong if he had called).

The Second Case: Does Egoism Give the Right Reasons?

Here is a case that could be viewed as presenting us with a variation on the same problem. Hitler’s actions resulted, among other things, in the deaths of several million Jews. In all likelihood, those actions were not in his interests. They did not help to promote, but interfered with, his having as rich and fulfilling and successful a life as he could have had. So, in this case, egoism gives the right answer about what he should have done. He should not have made the plans, taken the actions, or cultivated the obsessive hatred that led to those deaths. Acting as he in fact did made him worse off.

But surely, though this is the right answer, egoism gets it for the wrong reason. What Hitler did was wrong, not primarily because it did not promote his own interests but because it trampled on the interests and violated the rights of his victims. Hitler’s moral failure was not that he was imprudent or that he miscalculated.

The Third Case: Is Murder Acceptable for Trivial Gains?

We can push this further by considering counterfactual cases where imprudence or miscalculation do not figure. Ordinary moral beliefs have implications not just for cases that actually occur but also for cases that have not occurred. (In fact, they have to, in order to guide choice and action.) For example, almost everybody would hold that you shouldn’t be willing to kill someone if it happens to slightly advance your own interests. But imagine someone who could gain slightly by killing some innocent person. Suppose he would gain one extra dollar (or something worth a dollar to him), free and clear, after considering all costs and consequences of the act and that he has no other option that gains as much. Then, the egoist would have to say (at least) that the killing is permissible and possibly that it is morally required. He would have to say the same even if the cost in others’ lives were much higher – if the gain required the death of a dozen or a hundred or a million innocent people. (I discuss this argument at greater length in Killing for Trivial Gains.)

This, I submit, is morally outrageous, but it is a fair conclusion to draw if egoism is correct. The egoist can have no principled objection to killing others, no matter how many, if it leads to otherwise unattainable gains for himself, no matter how small. If, like me, and, I am confident, like most people on this list, you reject the notion that small gains can justify murder, I suggest that the right thing to do is to reject the moral theory, egoism, that leads to that conclusion.

I know that Objectivists are likely to view this argument with suspicion. Typically, they will offer either or both of the following objections. First, they will claim that the kind of case described cannot come up because there are no conflicts between rational interests (in a rational society). Second, they will claim that there is something wrong in general with basing conclusions upon arguments from hypothetical or counterfactual cases. To do so is said to be “rationalistic� and “not grounded in reality.�

The first objection can be treated briefly. The premise upon which it rests is a bit of unsupported dogma. There are no good arguments that rational interests never conflict. The best arguments in the neighborhood of that conclusion only show that conflict of interests is relatively rare, that interests are more typically harmonious than conflicting. I have discussed this in the paper on Rand’s ethics, and there is also worthwhile material in Antony Flew’s “Selfishness and the Unintended Consequences of Intended Action� in The Philosophic Thought of Ayn Rand.

About the second objection, I shall say a little more. The first point to bear in mind is that every moral principle has counterfactual implications – implications for what would be right or wrong in situations that do not occur or have not actually occurred. In fact, moral principles must have counterfactual implications if they are to guide choice, for they tell you what to do in situations in which you, at least, have never been before. Even something so mundane as “don’t cheat your employer� implies things like “if you were to be employed by IBM, then you shouldn’t cheat IBM� – and that holds true whether or not there is any prospect that you ever will be employed by IBM. It is perfectly legitimate, in general, to test a moral principle by asking whether its counterfactual implications are acceptable. If they are not, that’s a good reason for rejecting the principle.

Simply rejecting arguments from counterfactuals is not a serious option. We need them in order to apply any moral principles. The person who rejects some and accepts others looks suspiciously like he is arbitrarily accepting such arguments when they lead to conclusions he likes but rejecting them when they lead to conclusions he doesn’t like. There is nothing in general wrong with such arguments: if something is wrong with one of them, it is something other than just the fact that it begins from a premise incorporating a counter-factual. (Any Objectivists inclined to disagree are invited to think long and hard about the status of arguments about indestructible robots.) In any case, Michael Huemer has done a nice job of defending this type of argument and applying it to a case of murder for trivial gain at http://www.rci.rutgers.edu/~owl/rand.html. Take a look at section 5.3.

I won’t pursue this further except to state the obvious: It is far more certain that it is wrong to commit murder for trivial gains than that egoism is a correct moral theory. Since there is a conflict, egoism is the one to give up.

The Fourth Case: When is Honesty in Your Interests?

We need not, however, consider extreme counterfactual situations, where it is doubtful either that they actually occur or that, if they do, they can be known to occur, to find conflict between ordinary moral convictions and egoism.

Consider the importance of long-run consequences of action for the egoist. Any remotely credible version of egoism will insist that the interests by which an agent should guide his behavior must take into account the long term. Any egoism worth the name will allow that egoists may and often must rationally incur immediate costs or accept smaller immediate benefits than are available from some alternative course of action because refusing to shoulder the immediate costs or opting for larger immediate benefits will involve greater future costs or reduced future benefits that outweigh any immediate gains.

This attention to the long run is what makes it credible that egoists will be trustworthy, will keep their promises, pay their bills and so on. They can foresee that failing to be so or do so will have deleterious impact upon assets such as reputation and credit-worthiness that are worth more to them in the long run than any immediate gains they might secure by disappointing trust, breaking promises or ignoring bills.

Yet even here, in the domain in which egoists perhaps make the best and most persuasive case that their position leads to plausible moral conclusions, there are problems. For the fact is that the arguments work, insofar as they do, by appeal to the long run – but people do not always face a long run and sometimes, as in old age, know that they do not. Consider an egoist who, when young, adopts the following plan:

“While I am young, I have reason to consider the long term, including effects on my reputation and so on, so I will be honest, pay my bills on time, etc. But when I grow old, I will cheat people. I will run up credit card debts I don’t intend to pay. That way, I will live as well as people who follow the ordinary moral rules when I am young, but when I am old and therefore don’t expect to face the long term, I will live better than those who follow the ordinary rules. I’ll do at least as well for most of my life and better for part of it. By the time I am old, I expect to already have acquired and paid for most of the things a good credit rating is useful for. I don’t need to worry about severe personal consequences because, in the first place, since credit card debt is unsecured, banks will not be able to repossess anything I acquire in this way, and in the second, banks don’t send around kneecap-breakers to collect on bad loans. About the most I need to worry about is annoying phone calls – which I can stop by getting an unlisted number. No doubt there will be some bad consequences in terms of a slight deterioration in the fabric of trust on which a commercial civilization relies, but most of those will fall on others, not on me. As an egoist, I don’t see why I should be concerned.�

Once again, there is a conflict between a considered moral conviction – one that many egoists, certainly most here, share – and egoism. The ordinary moral rules do not say that it’s okay to stop being honest just because you won’t be around to suffer the bad effects of your dishonesty.

Some Non-Answers

In each of the four cases above, I have urged that egoism conflicts with moral convictions that, in fact, most self-described egoists share. A consistent response on the part of an egoist would be to reject those convictions. He could say that the conflict shows that at least one of the conflicting positions is mistaken, but that doesn’t settle which one it is. It is as consistent to hang on to egoism and reject the conflicting moral convictions as to hang on to the moral convictions and reject egoism.

True enough. That is a consistent position. But saying it is consistent is not the same as saying that it is reasonable or well-warranted. It might be reasonable to accept it if there were a powerful or compelling argument for egoism in the first place. That would be a reason for assuming that anything that conflicts with egoism must be mistaken. In fact, there is no such compelling argument, and therefore no presumption in favor of egoism. The egoist who takes this line is arbitrarily privileging one position over others which, on the face of it, are more plausible than that egoism is correct.

I think the above line of argument is correct. It really is more plausible that it is (say) wrong to commit murder for trivial gains than that egoism is correct. Still, that depends on judgments about the comparative plausibility of different views. I shall have something more direct to say later to those whose plausibility-judgments differ. For the present, however, it is interesting that the tack that most egoists have taken in response to such arguments amounts to an admission that I am right about the comparative plausibility of egoism and of the moral convictions to which I have appealed. The usual response has been that such convictions can, after all, be reconciled with egoism.

Several moves have been tried to effect the reconciliation (not all of which apply to all of the cases). It has been claimed, in one way or another, that egoism admits the relevance and correctness of the moral convictions in question. Five of the most common claims are (a) that there may be dramatic negative consequences for one’s interests attendant upon acting against these moral convictions, (b) that, because of our actual psychological make-up, we are happier if we accept and act upon the moral convictions in question, (c) that there are indirect (and outweighing) benefits to acting on these moral convictions, (d) that such purported conflicts overlook or underestimate the importance, for one’s interests, of acting on principle, and (e) that acting against these convictions will exact a toll – one that outweighs any advantages of doing so – on one’s self-respect. I do not think any of these succeed.

The Possibility-of-Disaster Move

Sometimes egoists point to the possibility of disaster that may be associated with acting in the kinds of ways that I have claimed are self-interested. The elderly cheater, for example, may live much longer than he expects (and therefore may face a longer term than his calculations were predicated upon) or may suddenly find himself afflicted with a potentially fatal but curable disease for which he cannot get the (expensive) required medical treatment without a good credit rating, etc.

All of this is true, but the response is more bluster than substance. Of course, disaster may result from executing the elderly cheater’s plans – but that is just as much true if he plans or acts differently. The fact that a course of action may turn out badly does not show that it was not rational or self-interested to adopt that course of action. One may also be killed driving to work, but that doesn’t show that it is not rational or self-interested to do that. The relevant question is how the possibilities of disaster, discounted by their probabilities, compare to the possibilities of gain discounted by their probabilities. It can be in a person’s interests to seek out even a small, highly probable, gain at the risk of a large but improbable loss. To plug in some numbers for the sake of illustration, suppose that the (essentially certain) gain from taking cash advances (and not repaying them) on all the credit cards a person with a good credit rating would possess is a hundred thousand dollars. If his risk of, say, a disaster that would cost him a million dollars is less than ten percent, then it’s rational for him to take the money (assuming that self-interest is the only thing that counts).

To say that the possibility of disaster shows he shouldn’t take the risk proves too much. It ignores the fact that, no matter what he does, there is a possibility of disaster – and, therefore, if it were correct, would show that he shouldn’t do anything. (Of course, that could have disastrous consequences, too!)

The Psychological Gambit

Some egoists claim that the reasons for having the kind of concern for others that would be exhibited in life-long honesty or in taking the trouble to make a phone call to get help for Kitty Genovese have to do with features of our psychological make-up. We are so constituted that we are happier if we have and act on those concerns than if we do not.

I see two problems with this. The first and lesser is that it makes our obligations in cases like these depend on our actual desires and motivations. If we don’t have the relevant motivations, then we also don’t have the obligations. If someone credibly denies having the motivations, there is literally nothing further for the egoist to say to him about why he, for example, shouldn’t cheat people in his old age.

But there’s a deeper problem, even if the case could be made out that everybody does have the relevant motivations. Saying that we are happier to have these kinds of concerns does not mean or imply that the concerns are self-interested. In general, we are happier (ceteris paribus) if we get or accomplish what we want, and we are frustrated or less happy if something interferes with our getting or accomplishing what we want. But that doesn’t show that the wants themselves are self-interested. If, for example, I want impoverished people to be fed, I will of course be pleased if they are. That doesn’t make wanting them to be fed a self-interested motivation. If anything, it shows the opposite – that I am not entirely motivated by self-interest. It’s not that I want people to be fed in order to be happy about it. Rather, I am happy about it because I really do want them to be fed. (This is one of the confusions involved in psychological egoism. Psychological egoists often say that some action is self-interested because of a normal consequence – namely, happiness – rather than attending to what the agent is aiming at. Happiness might be the normal consequence of wanting people to be fed and succeeding in getting them fed. That does not show the agent tried to feed them in order to be happy.)

Unless the egoist has some other reason for saying that the motivations in question are self-interested, the fact that acting upon them typically makes us happier does nothing to show that those motivations are self-interested. In fact, the point can be put more strongly: If the egoist is right that we are psychologically so constituted that we cannot be as happy without the kind of concern for others that would support acting on ordinary moral convictions and if he offers no further argument that concern of that kind is self-interested, what he is really doing is building a case that our basic motivations are not all self-interested.

The Indirect-Benefits Move

Another common move by egoists in the attempt to reconcile ordinary moral convictions with egoism is to appeal to indirect benefits of acting on those convictions. For example, it is said that we would all be better off if we could count on others intervening to help us (as in the Kitty Genovese case) when they can do so at small cost. Having the disposition to help is, so to speak, “insurance� – my contribution to living in a society in which people help (at small cost) and in which I may need such help. Similarly, there may be arguments about the indirect benefits to me of living in a society in which people are honest throughout their lives. (If people weren’t generally honest in such cases, then banks wouldn’t make loans to elderly people at all – and so on.)

It can be admitted that there may be some indirect benefits to me that flow from my own actions in such cases, but the question remains whether the benefits to me that flow from my own actions outweigh the costs to me that flow from those actions. In some cases – like the example of the elderly cheater – the slight unraveling of the social network of trust that will affect the cheater is overwhelmingly likely to affect his interests negatively less than the additional cash will affect those interests positively.

More broadly, the problem is that my compliance with a moral principle does not make others go along and do likewise nor does my non-compliance cause the non-compliance of others. The main benefits to me of the principle being followed come from other people following it, not from my following it. What would be best for me – even better than everyone following the principle – would be for everyone but me to follow it. That way I’d get the benefits of others’ compliance without enduring the costs of my own compliance.

[Note that this has the structure of a problem of public goods provision: All would be better off if all contributed, but each is better off if all others contribute but he doesn’t. In a pure case, each is better off not contributing no matter what the proportion of contributors is in the whole relevant population. In an impure case, in which there is some threshold of contributions above which the good will be provided and below which it will not, one would be better off contributing if and only if his contribution would make the difference between provision and non-provision. (Please do not read me as arguing for compulsory contribution to public goods. I am not.)]

The Importance-of-Principle Move

It is in our interests to have and act upon principles rather than to guide our behavior on a case-by-case calculation of costs and benefits. In fact, it can be argued that the position of the unprincipled case-by-case calculator is not really coherent. How, without any general principles, does he even recognize what a cost or benefit is? Principles save time and effort in calculating and reduce the risk of making mistakes when time is too short for careful calculation. It is true, of course, that adhering to such principles may also be costly in some situations, but, for well-founded principles, the (prospective) benefits of adherence outweigh the costs.

This, egoists sometimes say, is the self-interested reason for adherence to such principles as life-long honesty: Though there may be costs in particular cases, the long-run benefits are greater than the costs.

The most basic problem with this for an egoist anxious to reconcile his position with the kinds of moral convictions to which I have appealed above is that the person who rejects those convictions may also be acting on principle. An exploitative principle – such as that of the elderly cheater – is not therefore a non-principle. It was, in the story I told, adopted early in life and did not depend upon the agent making detailed case-by-case calculations.

It is sometimes objected that such a principle is not one that everyone could adopt (and therefore not really a principle?), but two points can be made in reply. First, there is no obvious reason that an egoist needs to accept the constraint that his principles be such that everyone could adopt them. The question he should be asking, as an egoist, is not whether everyone could adopt a principle, but whether it is to his own advantage to adopt it. The fact that it would be to the disadvantage of others or that it would be to his disadvantage if others adopted it is, in itself, no reason against his adopting it.

Second, even if, implausibly, the egoist has some reason for adopting only principles that everyone else could also adopt, a little tweaking will enable him to state such a principle. He could adopt a principle that says to everyone, “You may be a parasite just in case you rationally believe that you can get away with it for a normal expected life-span.� It is true that for the antecedent of that conditional to be satisfied, there would have to be others who are not parasites, but that doesn’t mean that it can’t be satisfied or that the exclusion-clause makes what the potential rational parasite adheres to “not really a principle.� It is neither as simple nor as morally attractive as one that forbids rational, principled parasites, but it is still a principle.

The Self-Respect Move

One final move made by some egoists is to say that our self-respect depends on our acceptance of and adherence to the moral convictions in question. This is really just a variant of the psychological gambit above and is subject to the same objections, but I mention it separately because I’ve seen it often.

For this special case, we can put the issue like this: Why will the egoist’s self-respect be injured by acting against or rejecting the moral convictions I have appealed to? There are two different things that might be understood by an appeal to the value or importance of self-respect and it’s important to see that one of them can’t possibly work. The one that can’t work fails because it involves circularity: To respect yourself is to think of yourself as deserving respect – but if you don’t have some other reason for thinking that, say, refraining from crime on principle is a good thing, then you wouldn’t have any basis for expecting to harvest self-respect for doing it. To put this the other way around, the clear-sighted criminal who is also an egoist would say that of course he could respect himself as much or more if he carries out self-interested crimes – after all, he’s just doing what his egoistic principle enjoins, pursuing his own interests. If you’re to convince him that he’s mistaken, you’ll have to point to some way the criminal acts harm his interests other than through their impact on his self-respect, because self-respect is a consequence of acting in accordance with one's convictions.

The other interpretation which might be suggested is that there are some psychological features of human beings, part of their nature, that place limits on the ways in which they can find self-respect. You might say that we have at least the rudiments of a moral code hard-wired into our brains. This may be true, but it leads to at least two deeper questions:

The first is whether everyone has the same (or a sufficiently similar) hard-wired “self-respect function.� Personal experience may be enough to tell you that you have such a self-respect function (though there are problems even there), but it surely cannot tell you that everyone is wired the same way – and therefore cannot provide any general arguments that egoists always have self-interested reasons to act in rights-respecting ways. Much more would be needed for that (and I don’t think it is likely to be forthcoming).

The second is a version of what I pointed out earlier about the psychological gambit in general: Is such a self-respect function even consistent with attributing self-interested motivation? Why is not the (alleged) fact that a person cannot respect himself if he engages in activity that would otherwise be in his interests – that is, without taking into account their impact upon his self-respect – a reason for thinking that he is not motivated entirely by self-interest? Or, to put the matter slightly differently, what makes it self-interested to do as one’s self-respect function dictates?

****

The above are the five main attempts that I know of to reconcile egoism with more ordinary moral convictions, but none is impressive. The Possibility-of-Disaster Move is completely worthless because it would rule out all action. The Importance-of-Principle Move is likewise worthless because it overlooks the possibility of exploitative principles. The Psychological Gambit and the Self-Respect Move both depend for their correctness on the availability of some other argument (which is not provided) for adherence to the questioned moral convictions. The best of these, the Indirect-Benefits Move, is simply empirically very weak.

In other words, the argument stands. Egoism in fact conflicts with moral convictions which are more plausible than it is that egoism is the correct moral theory. Therefore, egoism should be rejected.

The Third Problem: Egoism Undermining Itself

There is another way to show that egoism should be rejected. Remember that egoism is both a mandatory and universal theory. It addresses everyone and tells each person that he ought to be (not just that he may be) an egoist. Remember also that an egoist is someone who (at least) never acts against his interests. So, consider this possibility: Suppose it is possible to present an argument that a person can have a self-interested reason for acquiring some non-self-interested motivation, where a non-self-interested motivation is understood to be one that, when and if it is acted upon, requires the person so acting to knowingly act against his own interests. If such an argument can be provided, that will show that the person has self-interested reasons for ceasing to be an egoist. (I discuss this in a slightly different way in Egoism Undermining Itself.)

Can such an argument be presented? The short answer is Yes. But before doing this, I want to try to make one point clear because many people have found it confusing. I am not trying to argue something silly like “it’s in Jane’s interests to act against Jane’s interests.� The argument has to do with reasons a person may have for taking steps to bring about changes in her motivations. For a prosaic example, we might look at giving up a bad habit such as smoking. The smoker has reason in terms of his current motivations to give up his habit (more money, better health, longer life), but if he succeeds, his motivations themselves will change: He will no longer want to smoke. The motivations he has at one time give him a reason for taking steps to bring about a change in those motivations. Once he succeeds, his behavior will be guided by a somewhat different set of motivations. To return to the case of egoism, the argument will be that an egoist can have self-interested reasons at one time to take steps that will result in a change in her motivations such that she is, after the change, no longer an egoist.

How would the argument go? I’d like to lead up to it slowly. Imagine that you’re a universal egalitarian altruist in a world of other universal egalitarian altruists. In that world, each person acts only to promote the overall good. No one ever weights his own good more heavily than the good of others. And if the overall good appears to require personal sacrifice, each person willingly makes it. (You see, I’m not assuming any kind of insincerity or corruption.)

No doubt, you can give me half a dozen reasons that this wouldn’t be a very good way for the world to be, and so could I. I won’t try to list them but I’d just point out that these altruists, if they’re not stupid, could see these reasons, too. In effect, they’d be seeing that a world in which everybody tries to promote the overall good doesn’t achieve as high a score (in terms of overall good!) as a world where people have less universalistic, less egalitarian and less altruistic motivations. So, if their altruistic motivation isn’t just hard-wired into them, they would have reasons, in terms of altruism, to alter it – to become more concerned about their particular interests and less concerned about the overall good.

That is, they would have reasons at one time for doing something that would bring it about that at some later time they would act on different, more self-interested, motivations.

Now, you can probably see how the argument is going to go from here, but I’ll set it out with a concrete, though highly stylized, example:

You have a choice between living in two different communities. Apart from one fact, you’re indifferent between them. The one fact is that in one of the communities, there is a requirement that everybody take an “altruism pill�. It doesn’t alter any of your ordinary motivations. You’ll still be able to desire and pursue the very same things. It alters you in only one way. If you should ever happen to be in a situation where, by risking your life for other community members, you can save the lives of two or more of them, it will over-ride your ordinary motivations and you’ll do it. Now, assuming that the probability of accident, natural disaster, etc. is the same in both communities, and assuming that it is in your interests to live longer (other things being equal), you should join the community where you will have to take the altruism pill. It will result in an increased life expectancy for you since you are at least twice as likely to have your life saved as to lose your life saving someone else. That is, you can have self-interested reasons at one time for doing something that will bring it about that, at some later time, you will have some motivation other than or additional to self-interest.

Of course, there aren’t any real-world altruism pills. But speaking of magic pills is not essential to the main point: they are just a dramatic touch standing for something that is unquestionably real, that people can take steps to bring about changes in their motivations – through practice, training, acquisition of habits and various kinds of therapy. In principle, people can have self-interested reasons for ceasing to be entirely self-interested – that is, for ceasing to be egoists.

Is this in-principle result actual? I think there’s no serious room for doubt that it sometimes is. For example, the inculcation of military discipline in basic training is a close real-world analogue. A soldier undergoes training that is designed to bring it about that the soldier will obey orders even at great personal risk. His chances of surviving, under battle conditions, are better if he is a member of a well-disciplined group of soldiers, but being well-disciplined means that he is willing to knowingly act against his own survival. Less dramatic examples can also be given. It is not only possible in principle that people can have self-interested reasons for acquiring non-self-interested motivations: it actually occurs.

What does this mean with respect to the correctness of egoism? It means that egoism is false. Since egoism tells everyone that he ought to act in his self-interest, it tells the person faced with such a situation that he ought to take the steps that will result in changing his motivations. Once his motivations have changed, he is no longer an egoist. But since egoism is a mandatory and universal theory, the existence of even one person who ought not to be an egoist shows that the theory is false. The argument is simple: If egoism is true, then it is false. Therefore, it is false.

Follow-up

Will Wilkinson responded as follows. After that is my response to him.

Having read Rob Bass’s criticisms of rational egoism (at http://oocities.com/amosapient/egoism.html), I find that I have a few things to say in reply.

I am not concerned so much with the specific arguments that Bass advances; they are the same hackneyed arguments of which everyone interested in egoism has been long aware. One can only react to these arguments and the feeling of deja vu they engender with a heavy and wistful sigh.

I am concerned not with these old chestnuts but rather with the method of moral philosophy that permits these kinds of arguments to be made (again and again). This is the method of most of contemporary analytical moral philosophy. Bass shows that he has assimilated well to the prevalent mode of philosophizing, and this surely bodes well for him career-wise. Nonetheless, it is a very bad way of doing philosophy, as I shall argue.

The method of modern moral philosophy is, inter alia, anti-systematic, intuitionistic, and a prioristic. Some of these charges amount to about the same thing. To see that these charges apply to Bass’s method, let us see how he asks us to decide a dispute between contending ethical theories.

Bass writes that,

In ethics, as elsewhere, we have to start where we are. And “where we are� includes our possession of moral beliefs that we would not easily abandon and our engagement in practices that presuppose the correctness of moral beliefs. There still might be an argument in favor of some moral position that proceeds from our considered moral beliefs combined with other things we know to the conclusion that a given theory best harmonizes those prior moral beliefs with one another and with other relevant knowledge. Such a procedure is not necessarily bound to what we already happen to believe, because there may be mutual modifications of prior beliefs in light of the best theory and of the best theory in light of prior beliefs. Prior beliefs constrain what an acceptable theory will look like, and the best theory we can construct may provide a reason for rejecting or revising prior beliefs.

This is a decent description of John Rawls’s method of “reflective equilibrium.� The method is anti-systematic in that it just starts at the level of ethical convictions as if these were first principles and an appropriate basis for theorizing. It is intutionistic in that “considered beliefs� are granted a degree of justification just on the strength of our convictions in them. And it is a prioristic in that “considered beliefs� are taken to have some degree of epistemic justification independent of experience.

Furthermore, it is profoundly conservative. Bass acknowledges the method’s inherent conservativism when he remarks that such a method “is not necessarily bound to what we already happen to believe,� which implies that, on balance, the method does indeed tend to put you back right where you started, even if it doesn’t always do so. This is, of course, troublesome. But the real problem with reflective equilibrium, besides its mistaken coherentist conception of justification, is that the nature of the constraint of “considered� belief on theory is wholly unclear, and is poorly motivated to boot.

A considered belief is, more or less, a belief that one would be loathe to reject. It is a belief that one is really confident in. Beliefs that a person holds dear, for whatever autobiographical reasons, philosophers like to call “intuitions�. They are beliefs, as Bass says, that “we would not easily abandon.�

Now, it is completely mysterious why intuitions should constrain our theorizing. For to use these beliefs as checks on theory is to confuse the psychological status of a belief with its epistemic status. Confidence is a psychological, not an epistemic category. Probability in relation to a body of evidence is an epistemic, not a psychological category. Probability and confidence often wildly co-vary. Many religous people have a high degree of confidence in the existence of God, although the probablity of God’s existence relative to their evidence almost certainly approaches zero. Likewise, the theory of natural selection is highly probable relative to many of these same people’s evidence, yet they are not confident in it.

So, either using “considered beliefs� as a check on theory is a kind of category mistake, or “considered beliefs� have independent epistemic credentials, in which case they are are not a class distinct from theoretical beliefs, but are simply another set of theoretical beliefs with their own probabilities relative to one’s body of evidence. One’s set of prior beliefs just is a theory, and elements of that set are either confirmed or disconfirmed by experience, just like any theory. That one is comfortable with these beliefs, or that they got there first, does not say anything in their favor.

Now, how do we get evidence for any proposition whatever? From experience. Ethical propositions are not special in this regard.

In order to defend a position in moral philosophy, one has to first identify which empirical properties the moral properties are. In other words, one must identify the referents of moral concepts. Only when we know the referents of our moral terms can we know the truth conditions of our moral propositions. And we cannot assess the truth or falsity of a moral claim unless we know what the truth conditions for the claim are (i.e., until we know what fact about the world would make the claim true or false). This is just to say that before we can do moral epistemology, we have to do moral ontology and moral semantics.

A genuine refutation of Objectivist egoism must proceed by demonstrating that Objectivism has misidentified which empirical properties the moral properties are by using a naturalistic, a posterioristic, non-intution-mongering methodolgy.

If the real point of contention is the truth of Objectivist naturalism as a position in epistemology, then the debate should be refocused there. I think I could argue with Bass until the cows come home about the specific arguments and cases he presents – on the logical level he presents them. But this would be a vain and wholly imprudent pursuit. It is a waste of time debating a derivative issue when the participants in the debate do not agree on the logically prior matter of what it would take for one position to be shown superior to another.

-- Will Wilkinson

My response:

Will Wilkinson says that my objections to egoism (at http://oocities.com/amosapient/egoism.html) are “hackneyed� and “old chestnuts,� that anyone interested in egoism “has been long aware� of them, and that he can react only “with a heavy and wistful sigh.�

Perhaps so. I don’t claim they’re especially original. I think there’s a better way to describe it, though. It is a perennial temptation to suppose one can get the undoubtedly powerful engine of self-interest to pull the moral train, and, in every generation, some imagine they’ve made that remarkable discovery. Fortunately, however, in every generation, clearer heads have seen that nothing really new was being said and that the same old counter-arguments are still decisive. That’s why the same arguments get repeated. No new refutation of egoism is needed, just attention to the fact that it has already been refuted.

But Will disclaims interest (for the present) in directly addressing the arguments and says he will focus on methodology instead. I suppose his idea is that if he can show that the critics’ methodology is flawed, that will cast doubt on their results.

Yet there is something curiously blinkered about his attention to this issue. First, he addresses almost all his remarks to part of the first section of my essay and says almost nothing about the rest. Indeed, from some of his remarks, one might wonder if he had even read the rest. Second, what he actually says about the part he does discuss is almost invariably a misreading.

For some particulars, consider the following. After quoting some of what I had said on proper method in moral theory, Will remarks that: “[t]his is a decent description of John Rawls’s method of ‘reflective equilibrium.’ � He thinks that

the real problem with reflective equilibrium, besides its mistaken coherentist conception of justification, is that the nature of the constraint of “considered� belief on theory is wholly unclear, and is poorly motivated to boot.

I’ll try to explain some of the reasons for the constraint of considered belief on theory a bit below, and I have elsewhere discussed the coherentist conception of justification (at http://oocities.com/amosapient/coherence.html), but for the present, I’ll mention that there’s an even more distinguished practitioner and proponent of the approach than Rawls:

Here, as in all other cases, we must set down the appearances (phainomena) and, first working through the puzzles (diaporesantas), in this way go on to show, if possible, the truth of all the beliefs we hold (ta endoxa) about these experiences; and, if this is not possible, the truth of the greatest number and the most authoritative. For if the difficulties are resolved and the beliefs (endoxa) are left in place, we will have done enough showing.

That’s Martha Nussbaum’s translation of Aristotle’s NE 1145b1ff. in her essay, “Saving Aristotle’s Appearances� (in The Fragility of Goodness). Now, I wouldn’t want to suggest that Aristotle couldn’t be wrong, but perhaps Objectivists, who respect him as a great philosopher, should think twice before they reject as fundamentally confused the methodology that he employs “in all ... cases.�

Leaving aside questions of provenance, I’ll turn to Will’s objections:

The method is anti-systematic in that it just starts at the level of ethical convictions as if these were first principles and an appropriate basis for theorizing. It is intutionistic in that “considered beliefs� are granted a degree of justification just on the strength of our convictions in them. And it is a prioristic in that “considered beliefs� are taken to have some degree of epistemic justification independent of experience.

I think each of these is either false or misleading.

1. The method is not anti-systematic. It aims at developing a coherent and systematic moral theory. Perhaps Wilkinson has some strange notion of the importance of being systematic that requires that you start with a systematic theory rather than, if things go well, that you end up with one.

2. It does not treat ethical convictions as “first principles� if that means treating them as starting points that are beyond revision or reconsideration.

3. It is not intuitionistic in the sense he says. Considered moral convictions are not regarded as justified just because we feel strongly about them. A considered belief is not, as he says, just “a belief that one is really confident in.� First, we were talking about considered beliefs – beliefs that have survived reflection and criticism. For example, a person who just happens to believe that we should never lie probably does not count as having a considered moral conviction on the subject. To have a considered belief on that subject, he would have to have to have thought about things like inquiring murderers, what to do when telling the truth will be misleading and so on. Just happening to believe something without having faced or at least having been willing to face questions about how it fits in with other knowledge is not to have a considered belief.

There’s a second point, too. Suppose we are talking just about beliefs that we happen to have. (This may be an earlier stage in the process.) We still have to start there – because there isn’t any alternative. To believe something is to believe it true. Does he suppose it’s better to start with things we do not think are true? One might say that proceeding in the way I do requires a sort of theoretical assumption, but it is, in the first place, quite minimal, and in the second, it is one that neither Wilkinson nor any other moral theorist can reject, on pain of condemning his own theorizing. It is the assumption that we are not hopelessly bad at dealing with moral issues: having a moral belief or following out a course of moral reasoning is not just randomly related to being correct. If we don’t assume something like that, then we would be unwise to accord any credibility to moral beliefs that we happen to have – but equally, if we don’t assume something like that, it is deeply unclear how any elaborations of moral theory could help us make any progress.

4. It is not aprioristic. It does not assume that ethical convictions are justified apart from or independently of experience. It only recognizes something that should be obvious anyway: that we can have beliefs that do have a reliable connection to our experience but for which we are not currently in a position to articulate the connection or the reasons that it is reliable. (The ordinary non-philosopher really does know that there’s a world outside his own mind, even if he doesn’t know how to answer a skeptic.)

After these points, Will continues:

Furthermore, it [the method] is profoundly conservative. Bass acknowledges the method’s inherent conservativism when he remarks that such a method “is not necessarily bound to what we already happen to believe,� which implies that, on balance, the method does indeed tend to put you back right where you started, even if it doesn’t always do so.

This is bizarre. He adds an emphasis that wasn’t present in the text he was quoting, extracts an “implication� that was no part of what I meant, and calls it my “acknowledgement� of the conservatism of the method. For the record, I acknowledged no such thing. It’s possible that applying this method could leave a person just where he started, but it’s hardly likely. Almost always, he would not end up just where he started.

Near the end of his discussion, Will says:

A genuine refutation of Objectivist egoism must proceed by demonstrating that Objectivism has misidentified which empirical properties the moral properties are by using a naturalistic, a posterioristic, non-intution mongering methodolgy.

This is one of the places in which the oddness of Will’s focus only upon a small portion of my essay is evident. He speaks as if there were some Objectivist argument for egoism on the table that is at least a serious candidate to be considered a correct moral theory, but which I have somehow ignored or failed to address.

But surely, from reading the first section, he could have gleaned the following two facts:

1) I believe that no powerful or compelling argument has ever been offered for egoism (Objectivist or otherwise). It is no more serious a candidate than lots of other moral theories, and it has no prima facie credibility that should protect it in the face of obvious counter-examples (such as the wrongness of committing murder for trivial gains).

2) Even if he didn’t have a chance to read it, he should know that I have directly criticized the Objectivist arguments in a piece I have referred list-members to, The Rights (and Wrongs) of Ayn Rand.  I have not ignored or failed to address the Objectivist case for egoism. It is, in my judgment, flatly worthless. (This is not saying that there is nothing good or insightful in it, just that, as a case for egoism, the arguments do not stand up.)

Additionally, from reading the third part of my essay, Will should know that I presented an argument for the falsity of egoism that did not depend on acceptance of ordinary moral convictions. Even if every argument I made in the second section were mistaken (which they were not), I offered an independent argument that, given certain facts about human motivation and the kinds of reasons that people can and sometimes do have for altering their motivations, egoism could not be true because it issues contradictory prescriptions. It prescribes that everyone should be an egoist and also that some should not be.

I take it to be a minimal adequacy condition for the correctness of a moral theory that it not issue contradictory prescriptions. Egoism fails the test.

Rob Bass

Comments? I'd love to hear!

To Objectivism: Assorted Commentary

To Other Stuff