Utilitarianism, Philanthropy and the Act/Omission Distinction

By Mike Dorf

Professor Buchanan's post on Friday problematized a recent NY Times column by Peter Singer that I also found problematic--but my chief difficulty with Singer's analysis is somewhat different.  I'll try to spell it out here.

To review, Singer argued that it is possible for a philanthropist to use cost-benefit analysis to figure out where his or her charitable contributions can do the most good.  For illustrative purposes, he asked whether it would make more sense to donate money to fund a new wing of an art museum or to curing trachoma--and thus preventing blindness--in the developing world.  As Professor Buchanan notes in a post that is generally devoted to showing the dependency of cost-benefit analysis (and economic analysis more generally) on baseline assumptions, the question is really unanswerable, because institutions like art museums may help foster a moral culture that does more long-term good than any campaign focused on any specific illness.  I'm somewhat skeptical of the particular claim because it is possible to value art but not morals. See, e.g., Nazism.  Nonetheless, the broader point strikes me as right; tracing the good we do in the world is extraordinarily difficult so that we might say it is ex ante unknowable.

But let's put that point aside.  Suppose you are quite confident that Singer is right about the particular art museum.  That is, you have very good reason to think that people who go to enjoy the art in the new wing of the museum will not be moved to act in a morally better fashion than they would have acted if they had not gone to the art museum or, as in Singer's hypothetical example, if they had gone to see the same art but in the old display rather than in the new wing that donations will go towards constructing. Now is it true that your donation to fighting trachoma will do more good than your donation to building the new wing of the art museum?  I think the answer is yes but I have serious doubts about the mechanism Singer uses to get us there.

Singer makes some assumptions about how much it would cost to build the museum's new wing and to deliver aid to trachoma victims to propose the following thought experiment: 
[I]magine that an evil demon declares that out of every 100 people who see the new wing, he will choose one, at random, and inflict 15 years of blindness on that person. Would you still visit the new wing? You’d have to be nuts. Even if the evil demon blinded only one person in every 1,000, in my judgment, and I bet in yours, seeing the new wing still would not be worth the risk.
And so, if you yourself wouldn't choose to patronize the museum at that risk of blindness, then you shouldn't fund the museum rather than the trachoma treatment at the statistically equivalent risk of failing to cure blindness.

One standard objection to utilitarianism is that it falsely assumes an ability to make interpersonal comparisons of utility.  In this example, that objection would be that while any individual might prefer avoiding the risk of blindness to seeing the art in the new wing, the actual choice is whether to provide an enhanced experience of art to one set of people or to fail to intervene to prevent blindness for a different set of people.  If we cannot make interpersonal comparisons of utility then we cannot know that the additional pleasure for the hundred (or thousand) people who see the art in the new wing does not outweigh the additional discomfort for the one other person who experiences the blindness.

I don't buy the objection to interpersonal comparisons of utility.  Just about all systems of morality assume its possibility.  Think of the Golden Rule, for example.  Moreover, biology appears to confirm that we (and other animals) have mirror neurons that enable us to understand the world as seen by others.

But there is, in my view, a more profound objection to utilitarianism: It does not respect the distinction between acts and omissions.  In Singer's demon hypothetical, visiting the new wing of the art museum inflicts blindness (via the demon's act), but in the original question, donating to fund the new wing of the art museum means failing to act to cure trachoma.  The difference is often critical.

Consider this hypothetical example.  Suppose you wake up on your seventieth birthday to discover that you have "Buchanan-Dorf Syndrome", a degenerative disease of the optical nerve that, if left untreated, will leave you blind by the time you are seventy-five.  Given your current life expectancy, that means you will experience about fifteen years of blindness.  Your doctor tells you that there is a very promising treatment available for Buchanan-Dorf but it has not yet been approved for use in the U.S. and therefore your health insurance won't cover it.  You can travel to a foreign country for the treatment at a cost of $100,000.  Or (per Singer's numbers), you can decide to suffer with Buchanan-Dorf and donate your $100,000 to spare one thousand people with trachoma from fifteen years of blindness each.  Would it be wrong to spend the $100,000 on yourself?

Deontologists have little difficulty with such a question.  They say of course it's not wrong.  It would be very generous of you to suffer the blindness so as to spare a substantially larger number of people blindness, but you are under no moral obligation to sacrifice your own well-being for others. Indeed, you are not even under a moral obligation to spend your spare $100,000 on curing the trachoma of strangers rather than on a super-luxury car (like this one).  That is why we do not ordinarily think of people who buy luxury items as killing or blinding the people whom they might have aided if they had instead donated their money to charity.

What would Singer say about my Buchanan-Dorf syndrome example?  Unlike in the case of the art museum--where we are in some sense comparing apples (art) to oranges (vision), here we have an apples-to-apples (vision-to-vision) comparison.  Singer would surely say that you would do more good by funding trachoma treatment for a thousand people than by funding Buchanan-Dorf treatment for yourself.  But does that mean that you are obligated to forgo treatment for yourself?

Here I think that Singer's utilitarianism gets stuck.  Singer could say that you are not obligated to forgo the treatment for yourself by denying that his utilitarianism has anything to say about moral duty.  Note that his Times essay talks about doing good but not at all about right, wrong, morality or immorality. 

But if that's a general feature of utilitarianism, then utilitarianism is monstrous.  Suppose that the situation were reversed.  Suppose that the eyes of persons with a harmless condition called Dorf-Buchanan Syndrome contain a chemical that, even in tiny quantities, provides the only known cure for trachoma, so that cutting out your eyes and grinding them up will save a thousand people from blindness for fifteen years.  If utilitarianism has nothing to say about right or wrong, then Singer cannot even say that it would be wrong for the state (or some third party) to forcibly remove your eyes to grind them up.

The foregoing example is simply a variant of a very old challenge to utilitarianism.  It can take the form of the trolley problem or the question whether a doctor may cut up a healthy person to harvest his organs and thereby save five unrelated sick people.  For thoroughgoing utilitarians the answer is always given by numbers: Which combination of acts and omissions leads to the greatest good for the greatest number?  For deontologists, the distinction between acts and omissions is crucial.

Utilitarians, for their part, have posed challenging hypothetical examples for deontologists--mostly dealing with catastrophes.  If it is always wrong to torture, the utilitarian asks, does that mean that it is wrong to torture the terrorist who has planted the bomb that will kill a million people?

In response to these various challenges, utilitarians and deontologists have developed variations that blur the lines between the two.  We have utilitarians with side constraints and we have rule utilitarians who purport to be able to derive most of the rules of deontological moral reasoning from considering long-term as well as short-term consequences of classes of actions.  From the other side, we have threshold deontologists, whose deontology gives way to utilitarianism upon the crossing of some threshold from mere ordinary harm to catastrophic harm.

I'm not going to defend any of these comprehensive positions here.  Instead, I simply want to make an observation about how I end up skeptical of pure utilitarianism from the opposite direction than Professor Buchanan does.  As he rightly notes, utilitarianism is frequently used by right-leaning economists as the ground for resisting redistributive policies: If we aggregate all utility, then we needn't worry about distributional questions, and the redistribution process itself will have transaction costs that lower total welfare, so redistribution is wrong, they say.

But I tend to think that utilitarianism, used honestly, has the opposite problem.  Given the diminishing marginal utility of wealth, in a wide range of circumstances one can do more good with one's money by giving nearly all of it away than by spending it on oneself and one's family, even once one takes account of transaction costs.  So if the greatest good for the greatest number really is a moral principle, then one is obligated to give away nearly all of one's money--for otherwise one is effectively choosing creature comforts over blinding people and inflicting starvation on them.  In other words, the real problem with utilitarianism is that it is too demanding, not that it is insufficiently demanding.