Why Might the Parties Choose Utilitarianism?

I began by summarizing a section of the book that I did not ask you to read (we are reading too much of this extremely dense material as it is). This section describes a chain of reasoning that might lead the parties in the original position to choose utilitarianism.

In the parts we did read, Rawls argued that they would have decisive reasons not to follow this chain of reasoning and so they have decisive reasons to reject utilitarianism.

The parties in the original position would choose utilitarianism if they based their decision on maximizing their expected utility, the product of the probability of an outcome and its value; see the handout for an example of how decisions based on maximizing expected utility work.

I illustrated the point with an example involving self-driving cars. Suppose a self-driving car comes on a situation where it is going to kill five people standing in the road and the only alternative is to crash the car, possibly killing the one person inside the car. We can program the car to save the person in the car or to save the greatest number.

If you were in the original position you would choose the program that saves the greatest number. That is because, in the original position, you are trying to do as well for yourself as possible and you don’t know who you are in the real world. In the case at hand, you are five times more likely to be one of the people in the road than you are to be the person inside the car. So saving the greatest number has the higher expected utility for you.

The case for thinking the parties would choose utilitarianism rests on treating the choice of rules to govern a society as being similar to the choice of rules to govern a self-driving car. Utilitarians think this is obviously true while Rawls thinks it is false.

Probability

Rawls believes that the parties in the original position would choose to maximize average utility only if two conditions are met:

1. It is rational for them to maximize their expected utility (rather than following the maximin rule).
2. They can assign probabilities to the possible outcomes of their choices.

Rawls’s chief reason for denying that this makes sense is the familiar one: maximizing expected utility is too risky for this particular choice. Instead, the sensible choice is to follow the maximin rule.

If the parties were engaged in an activity where there would be repeated plays and no particular loss would be devastating, like low stakes gambling, it would make sense for them to maximize expected utility. No loss would wipe them out and they will come out ahead in the long run. But the parties in the original position have to make a single decision that will never be repeated and that could have calamitous implications over the course of their entire lives. Consequently, Rawls reasons, it makes no sense to take the riskier rather than the safer option.

He added an argument to the effect that the parties are incapable of estimating probabilities; this is the second point above. This has been a perennial thorn in my side because I can’t get a handle on what they’re supposed to be incapable of estimating. I have come to the conclusion that the wording in A Theory of Justice is misleading and that the real idea is better expressed in a different publication.

The handout gives two passages from Rawls. The first is almost certainly wrong: the parties do know that their chances of being any particular person are equal to their chances of being any other person. The second makes sense, though. It says that the parties cannot estimate the probability of being in any particular circumstances. So if they choose rules that allow slavery in their society, they do not know how likely it is that they will wind up as slaves. The risk could be very small or very large. Given that they do not know the probabilities, Rawls thinks it would be foolish of them to risk a social system like utilitarianism that could, conceivably, allow slavery.

Finality and Stability

In §29, Rawls advances two arguments that, in my opinion, boil down to one. These arguments appeal to what Rawls calls finality and stability. Finality means that the parties can only choose principles that are final: that was one of the conditions on the original position. Stability means that they can only choose principles that they would accept if they grew up in a society governed by them. That is also one of the conditions on the original position. (These conditions are listed in the handout on the original position.)

Rawls claims that these considerations favor his principles over utilitarianism because it is possible that some people would find life in a utilitarian society intolerable. If that happened, they would seek to change the society (contrary to the finality condition) and, of course, they would not accept its rules (contrary to the stability condition). By the rules governing the original position, the parties must avoid rules that would fail either condition, so they would reject utilitarianism. But the reason why a utilitarian society would fail the conditions is the same one Rawls had used before: someone in a utilitarian society could be a big loser and find life as a loser intolerable.

Is Rawls Cheating?

Helena posed an objection that utilitarians might make. It goes something like this. “Rawls says that the parties in the original position cannot follow the normal rules for making a rational decision when faced with uncertainty about what will happen, namely, by choosing the option with the highest expected utility. He says that this is so because he says they cannot estimate the probabilities of outcomes and, consequently, they cannot calculate the expected utility of the choices before them. But the only reason why they lack this information is because Rawls took it away from them. He designed the original position so they would not be able to estimate probabilities. That means that the original position does not pose a reasonable test of whether utilitarianism is preferable to Rawls’s principles of justice. It is set up so that utilitarianism will not be chosen.”

I think that this is a plausible objection for utilitarians to make. I think that Rawls would say something like this in response. “I deprived the parties in the original position of information about probabilities for good reason. If they had that information, their choices could easily be unfair. That is why I did not let them know about trivial things, like the ratio of right handed people to left handed people. If they knew that, they would set up the rules to favor the righties knowing that there is a very good chance that they will benefit from that. Since rules that favor righties over lefties would be unfair, they should be deprived of this kind of information. As it happens, once you design the original position so that its decisions are guaranteed to be fair, you discover that the parties would not choose utilitarianism. That is not cheating, it’s an interesting result!”

I’m going to leave it to you to decide for yourselves about who is right here.

Would the Parties Reject Utilitarianism?

Rawls’s chief argument is that utilitarianism is too risky. A society governed by utilitarian principles could possibly do something horrible like instating slavery. By contrast, a society governed by Rawls’s principles cannot do such a thing: it has to guarantee equal basic liberties for all, equal opportunity, and the difference principle.

Utilitarians should start by saying that they would not do anything horrible. That doesn’t make any sense if you’re trying to maximize happiness, after all. Who would ever think that slavery is a way of making people happy?

But suppose you pressed them with a real or hypothetical example in which enslaving some people really did bring about more utility overall than any available alternative to slavery.

Utilitarians would have to admit that it is at least possible that they would be willing to enslave some people. But, they would say, this would happen only in dire conditions, when life was bound to be intolerable for some people anyway. Suppose there is some catastrophe that can only be addressed by something like a military draft. If the alternative is catastrophe, and you said it was, then, a utilitarian will say, it only makes sense to institute the draft or whatever form of slavery is necessary. Why wouldn’t the parties in the original position see it that way too? They don’t want to be wiped out in a catastrophe, after all.

At this point, things become murky. Rawls tells us that he instructs the parties in the original position to come up with principles of justice for a society whose members are in what he calls the circumstances of justice. This means that they live in a world where material goods are only moderately scarce: people can live at peace with one another by obeying simple rules, much as Hume described. They are supposed to assume that they are not threatened with catastrophe that will obliterate their way of life.

Utilitarians will object that Rawls tries to make them look bad by describing what they would do in conditions of great scarcity or some other emergency. But his own theory explicitly assumes that there will not be great scarcity and that people will comply with the rules. The point is not that he is making unrealistic assumptions. It is that he develops his theory for a specific set of circumstances but criticizes utilitarians for what they would recommend in very different circumstances. This strikes utilitarians as an unbalanced comparison.

Rawls, of course, does not see it that way. He sees utilitarianism as being too reckless with individual lives. They are willing to sacrifice anyone in the right circumstances. Rawls thinks that the parties in the original position would not be willing to accept the risk of being the ones who are sacrificed.

Which is it? Again, I will let you decide.

Key concepts

There is really only one question: is utilitarianism is too risky for the parties in the original position or not? That is the whole ball of wax.