John Roemer is surely right that equality and responsibility are not opposed notions. How disappointing, then, that he ties responsibility to individual control and to the likelihood of similarly situated people behaving in similar ways. That approach strikes me as a dead end. It rests on the idea that ascriptions of responsibility are at odds with explanations in terms of social factors, and thus makes Roemer’s position vulnerable to those who claim an opposition between equality and responsibility.
Critics of the welfare state, as Roemer points out, sometimes doubt that it has any room for ideas of individual responsibility. Such critics often charge as well that liberals are too soft on crime. These criticisms have a common root. Both grow out of a concern that costs are being unfairly imposed on others — taxpayers in the one case, victims of crime in the other. Society as a whole is being charged with ensuring that individual lives go well, and that, according to the critics, is an unfair distribution of burdens.
Defenders of equality should try to answer these criticisms directly — to present a view of responsibility focused less on control, and more on the costs that each person’s choices impose on others. For example, egalitarians and their conservative critics agree that people with expensive tastes should have to forego other things in order to satisfy those tastes; such people are not entitled to more resources simply because they are so hard to satisfy. That point surely holds quite apart from any questions about how the expensive preferences were acquired. Whether their tastes came from choice, inheritance, or accident, people are presumed to be in a position either to modify the tastes, or to adapt their other desires to the resources to which they are entitled. The underlying idea is that it is not fair to shift the costs onto other people in society. So, for example, the fact that members of some groups are more likely to smoke than others cannot, by itself, show that they ought not to be held responsible for smoking’s effects. Not all influences that increase the probability of smoking count as burdens that society as a whole must shoulder.
The real question has to be whether each person has a reasonable chance at a decent life, in which they can adjust their tastes to their fair share of resources. Canadians and Americans live in societies in which many people lack such a chance, and making sure people have a fair share of resources and opportunities is of the utmost importance. But the reason is not that choice is impossible for people who lack a fair chance; the problem instead is that enduring such conditions is more than we can reasonably ask of anyone.
Roemer comes close to conceding this point when he says that society must decide which factors are and are not within a person’s control. The distinction between Roemer’s examples of a jaywalker and his cautious pedestrian isn’t a matter of control — the cautious pedestrian could have waited until there was no traffic coming from any direction, or stayed home altogether. Instead, the difference is that the jaywalker presumably took unreasonable risks. More generally, the boundaries of responsibility always depend on how hard we think someone in the circumstances ought to try. That in turn depends both on how difficult the task is, and what else is at stake.
Notwithstanding his suggestions of a quasi-scientific test for choice, Roemer’s view may well be that in setting the boundaries of individual responsibility, we must decide whether making individuals bear the costs of particular choices places an undue burden on them. But whether the burden is undue cannot be determined by statistical distributions alone. A person’s responsibility does not depend on whether others, similarly situated, have made similar choices. Where the enemies of equality go wrong is in supposing that people should be held responsible for everything, no matter how dreadful the choices facing them. That is a moral mistake about fairness, not a metaphysical error about who’s in control.
