Wednesday, June 28, 2006

Psychological Egoism Just Isn't True. And I'm Not Saying That Just 'Cause It's in my Self-Interest, Either

In a recent post, I commented that arguments for psychological egoism are laughably bad and that I thus wasn’t going to address them further. Tom, quite fairly, called me out on the claim, asking whether I could provide either a summary or a link to a good argument against PE. So in what follows, I’ll attempt to show why psychological egoism is a doctrine that ought to be rejected.

First, we should probably explain what psychological egoism actually is. Or rather, we’ll start with what it’s not. PE is not a normative theory; it is not a claim about how people ought to behave. Rather, PE is a descriptive claim, a claim about human psychology. Formally, PE maintains that, as a matter of fact, all human beings are always selfish. Proponents of PE have included Thrasymachus from Plato’s Republic, Thomas Hobbes, Jeremy Bentham and Ayn Rand. (Right now, I’m hearing the old Sesame Street song. You know. “One of these things just doesn’t belong here…”). The trouble, of course, is that on its surface, PE is just, well, false. Plenty of people do plenty of unselfish actions. I give money to charity. I help an old lady cross the street. I buy my son a book at the bookstore and pass on getting one for myself. I give my girlfriend the last glass of wine. All of these acts appear, on the surface, to be unselfish acts.

Of course, the PE has an explanation. One strategy to take is to offer what James Rachels has called the wants argument.[1] Here’s the argument:

  1. Everyone always does whatever he/she most wants to do.
  2. Always doing whatever one wants to do is selfish.
  3. Therefore people are always selfish.

This is indeed an argument for PE. It’s just a really crummy one. Indeed, it’s an example of the fallacy known as a definitional dodge, salvaging the argument by changing the meaning of a word (“selfish” in this case) into something not generally recognized as being what the word means. Consider: the claim that the PE is making here is that selfish just means “always doing whatever one wants to do.” Or in other words, because all of our actions have the general form

Miller most wants _________

it must be the case that all of our actions are selfish. Whatever action goes into the blank, be it “to save the starving” or “to steal from the poor,” the action will be selfish because, in either case, I will simply be doing whatever it is that I most want to do.

The problem, however, is that saying “Miller most wants X” is not what we generally mean by the word “selfish.” Rather, whether or not I am selfish depends upon what does into the blank space. The mere form of this statement does not make Miller selfish. Rather, the question of whether Miller is selfish depends upon what goes in the blank. If I want only my own happiness regardless of the consequences to others, then I am selfish. If I want other people to be well-off even at some cost to my own happiness, then I am unselfish. To say that everyone is selfish because we all do whatever we most want to do is to misuse the word “selfish.”

A second possible argument for PE is what Rachels calls the satisfaction argument. The form of the argument is pretty similar to the form of the wants argument:

  1. We always perform whatever action will give us the most satisfaction.
  2. Performing the action that maximizes our own satisfaction is selfish.
  3. Therefore we are always selfish.

As with the first argument, the difficulty here will be with (2). Consider the following example, cited in just about every intro-level paper on psychological egoism:

Mr. Lincoln once remarked to a fellow-passenger on an old-time mud-coach that all men were prompted by selfishness in doing good. His fellow-passenger was antagonizing this position when they were passing over a corduroy bridge that panned a slough. As they crossed this bridge they espied an old razorbacked sow on the bank making a terrible noise because her pigs had got into the slough and were in danger of drowning. As the old coach began to climb the hill, Mr. Lincoln jumped out, ran back, and lifted the little pigs out of the mud and water and placed them on the bank. When he returned, his companion remarked: “Now Abe, where does selfishness come in on this little episode?” Why bless your soul, Ed, that was the very essence of selfishness. I should have had no peace of mind all day had I gone on and left that suffering old sow worrying over those pigs. I did it to get peace of mind, don’t you see?”[2]

To see why this is a bad argument, let’s change the example slightly. Suppose pretty much the same example except that this time, I am in Lincoln’s place. This time, however, when I ask the driver to stop, I get out of the coach, walk over to the baby pigs, and hold their tiny heads under the mud until the struggles stop. I then walk back to the coach, grab a bat, and beat the mother pig to death. When you, looking rather appalled, ask me what the hell I was doing, I reply that I hate pigs, that I was afraid the mother might find a way to rescue the piglets and that even if she didn’t, the damn thing would just go have another litter of the bastards anyway; I just couldn’t have lived with myself had I passed up the opportunity to off a few of the fuckers. Now according to the PE, there is absolutely no morally significant difference between my drowning of the baby pigs and Lincoln’s rescuing of them. Nor would the story change if I substituted human infants for the baby pigs. In both cases, Lincoln and I simply act to satisfy our preferences.

The problem, once again, is that the fact that I satisfy my preferences isn’t by itself enough to show that I act selfishly. Rather, whether or not I am selfish depends on what kinds of preferences I actually have. To the extent that I get satisfaction from helping others, I am unselfish. To the extent that I fail to get satisfaction from helping others and instead get satisfaction only when I help myself, I am selfish. To put the point another way, to say that I get satisfaction from X is to say that I already had some preference for X or some desire to X. And it is the prior preference for or desire to X that is either selfish or unselfish. The fact that I act on that preference or desire is not itself evidence that I act selfishly.

Indeed, the very claim that humans always act to satisfy their preferences is itself incoherent. As Joel Feinberg points out, the very formulation of the theory leads to an infinite regress.[3] Consider:

“All men desire only satisfaction”
“Satisfaction of what?”
“Satisfaction of their desires.”
“Their desires for what?”
“Their desires for satisfaction.”
“Satisfaction of what?”
“Their desires.”
“For what?”
“For satisfaction.”

And so on, ad infinitum.

In short, the claim that people are always selfish because they always act to satisfy their preferences is (a) false, since it “selfish” means something entirely different, and (b) not even really coherent in the first place.

Perhaps, then, we could reformulate PE yet again. Maybe PE isn’t claiming that we’re all selfish. Perhaps instead the claim is just that we’re all self-interested. What’s the difference? We might define selfishness as

S: P acts selfishly in performing A iff A benefits P and A harms Q and P ignores the harm that A causes to Q.

Self-interest, on the other hand, we might define as

SI: P acts self-interestedly in performing A if A will benefit P.
To act self-interestedly, in other words, is to act in such a way as to benefit myself. Selfishness implies an additional step, one in which I actively harm (or at the very least, actively ignore the harm my action might cause) someone else. The reformulated version of PE, then, would hold that people always act self-interestedly. Unfortunately, this won’t do, either. It’s just plainly false. There are a number of activities that people perform that are not in their self-interest. I might ignore the signs that I have cancer. Or have sex with a prostitute without wearing a condom. Or ride a motorcycle without wearing a helmet. Or smoke two packs of cigarettes a day. All of these actions might well give me pleasure. Not a single one of them is in my self-interest, though.

There is, however, one final version of the theory. Let’s try it one more time:

PE: people always act in such a way as to maximize their expected utility.

This is probably the strongest version of PE that there is. It does, however, suffer from the problem that it looks to be empirically false. Plenty of people claim that they don’t do this sort of thing when they act. The typical PE response here is that people who claim not to act in such a way as to maximize their expected utility are really just deceiving themselves.

It’s hard to know where to start analyzing this baldly elitist claim. Do I begin with the arrogance of your claiming to somehow have a better understanding of what goes on in my head than I have? Tempting, but, elitist as it sounds, it may very well be the case that some experts have a better understanding of my psychology (or rather, my brain chemistry, but I’m speaking loosely here for a moment) than I have. So instead, I’ll point out that a defense of PE that rests upon the claim that most people deceive themselves regularly about their motives is, quite simply, no longer within the realm of empirical science. Alex Moseley explains the problem nicely:

On this point, psychological egoism’s validity turns on examining and analyzing moral motivation. But since motivation is inherently private and inaccessible to others (an agent could be lying to herself or to others about the original motive), the theory shifts from a theoretical description of human nature--one that can be put to observational testing--to an assumption about the inner workings of human nature: psychological egoism moves beyond the possibility of empirical verification and the possibility of empirical negation (since motives are private), and therefore it becomes what is termed a “closed theory.”

A closed theory is a theory that rejects competing theories on its own terms and is non-verifiable and non-falsifiable. If psychological egoism is reduced to an assumption concerning human nature and its hidden motives, then it follows that it is just as valid to hold a competing theory of human motivation such as psychological altruism.

Moseley goes on to offer a quick sketch of psychological altruism, or the view that all our actions are motivated by altruistic behavior. Like PE, PA cannot be refuted by empirical evidence, since, after all, any empirical evidence can be explained away as an act of self-deception. PA is also a closed theory, one that will in fact explain the world just as well as PE (since, again, nothing can really count as empirical evidence against either claim). In the end, our decision to side with PE rather than PA will be arbitrary; one might as well just flip a coin. A coin flip, however, is hardly an appealing way of determining the account of moral motivation that we’re going to use as the backdrop for the science of economics.

What does appear to be true is that humans are largely motivated by self-interest. That weak claim is enough to get economics off the ground. It does mean, however, that sometimes we will perform actions that are not motivated by our own self-interest. At times, we act altruistically. We’re just more complicated than the PE wants to admit.


[1] See James Rachels, The Elements of Moral Philosophy, 2nd ed, (McGraw-Hill, 1993).

[2] Quoted from the Springfield (Illinois) Monitor by F.C. Sharp in his Ethics (Appleton Century, 1928).

[3] Joel Feinberg, “Psychological Egoism,” from his Reason and Responsibility, 4th ed. (Wadsworth, 1978).

4 Comments:

Blogger Thomas said...

Joe, thanks for the quick tour of PE, PA, and SI. I agree that PE and PA are essentially equivalent concepts, and previously I had rejected both in favor of SI.

Now, I'm not so sure that PE and SI really are different. Consider your definitions of "selfishness" and "self-interest":

S: P acts selfishly in performing A iff A benefits P and A harms Q and P ignores the harm that A causes to Q.

SI: P acts self-interestedly in performing A if A will benefit P.


But S, as you define it, strikes me a special case of SI; e.g.:

P acts self-interestedly in performing A if A will benefit P, and in some instances A will benefit P iff A benefits P and A harms Q and P ignores the harm that A causes to Q.

I also could write it this way:

P acts self-interestedly in performing A if A will benefit P, and in some instances A will benefit P iff A benefits Q and A harms P and P ignores the harm that A causes to P.

Which seems also to make PA a special case of SI.

Here's the most general statement of SI, which encompasses SE, SA, and neutrality:

P acts self-interestedly in performing A if A will benefit P. In some instances A will benefit P iff A benefits P and A harms Q and P ignores the harm that A causes to Q. In some instances A will benefit P iff A benefits Q and A harms P and P ignores the harm that A causes to P. And in some instances A will benefit P regardless of its effects (if any) on Q.

Now, turning to your "strongest version" of PE, which is:

People always act in such a way as to maximize their expected utility.

Why is that PE instead of SI? Why not this?

P always acts self-interestedly in performing A if A will maximize P's expected utility. In some instances A will maximize P's expected utility iff A benefits P and A harms Q and P ignores the harm that A causes to Q. In some instances A will maximize P's expected utility iff A benefits Q and A harms P and P ignores the harm that A causes to P. And in some instances A will maximize P's expected utility regardless of its effects (if any) on Q.

So, it seems to me that self-interest is the key, and that self-interest can cut in any direction: toward "selfishness," toward "altruism," or toward indifference.

It's not necessary to "get into someone's head" to accept SI as the explanation for behavior, for SI encompasses all possible motivations. Anyway, if you can reject PE for its "elitist" mind-reading, why can't I reject your defense of interpersonal utility comparisons (two posts ago) for the same reason?

Next, I must address your view of self-interest. You say:

There are a number of activities that people perform that are not in their self-interest. I might ignore the signs that I have cancer. Or have sex with a prostitute without wearing a condom. Or ride a motorcycle without wearing a helmet. Or smoke two packs of cigarettes a day. All of these actions might well give me pleasure. Not a single one of them is in my self-interest, though.

You are working from a peculiar definition of self-interest; viz.:

Self-interest is defined not by how a person perceives his self-interest in the present but by (a) how that person might perceive his self-interest in the future when he reflects on the consequences of his past actions, or (b) how others might perceive that person's self-interest based on their own preferences (e.g., wealth or health maximization upon reaching old age).

Glen Whitman (Agoraphilia) had something to say about that, in the context of a post about "libertarian paternalism":

[T]he new paternalism blithely assumes that, when your present self can impose costs on your future self, the outcome is necessarily bad. But preventing harm to the future self might involve even greater harm to the present self. There’s no valid reason to assume, when there is an inconsistency between present and future interests, that the latter must trump the former.

The whole thing is well worth reading. Here's the link:

http://www.cato.org/pubs/pas/pa563.pdf

Toward the end you mention the "moral motivation that we’re going to use as the backdrop for the science of economics." The point (made in 1776 by Adam Smith) is that "moral motivation" is essentially irrelevant to economics. Self-interest explains economic behavior, and self-interest -- as I've tried to indicate here -- is a term that can encompass any kind of "moral motivation," which is another way of saying that "moral motivation" is irrelevant to economics.

10:11 PM  
Blogger Matt McIntosh said...

I basically agree with your main point, but there are some dubious details along the way.

1) Classing Rand as a proponent of PE is roughly equivalent to saying Christians don't believe in the Devil. Rand believed in altruism alright, she just thought it was the root of all evil.

2) In what is ostensibly a discussion of a descriptive theory, you're using the pig-killing example deliberately for *moral* shock value. This is low and you know it. You show that you know what you're doing when you say that "according to the PE, there is absolutely no morally significant difference between my drowning of the baby pigs and Lincoln’s rescuing of them." But this is wrong, if PE is a descriptive psychological theory: at most it says there's no *psychological* difference and is entirely silent on the matter of morality.

3) The "regress" argument only makes any sense if you think it's possible to have a preference for "satisfaction" in the abstract. The fact that this is logically incoherent should be a tipoff that something is wrong. Preference implies an object; satisfaction is an analytical relation between an object and a desiring subject. You can't have a preference for satisfaction by definition. Of course this is a definitional dodge, which makes it a pretty useless argument, but that's a different issue.

4) I think you probably also know you're eliding over a matter of significant philosophical debate with this: "All of these actions might well give me pleasure. Not a single one of them is in my self-interest, though." It really depends on how we slice and dice the self -- they might be in your short-run self-interest, though not your long-term self-interest. But is your self at t the same self as at time t+20? Etc. You could definitionally dodge your way out of this one too, though of course what you're left with wouldn't have much teeth.

5) It's also not clear to me what's "elitist" about the claim that people regularly and systematically decieve themselves -- in fact this happens to be true, just not in the way the PE people think.

12:08 AM  
Blogger Thomas said...

P.S. Regarding IUCs, check out this post and comment thread at Agoraphilia:

http://agoraphilia.blogspot.com/2006/06/iud-for-iucs.html

12:44 PM  
Blogger Scott said...

I agree with you.

But you're overplaying your hand: "laughably bad" is too strong---maybe far too strong.

3:34 PM  

Post a Comment

Links to this post:

Create a Link

<< Home