The social model of disability holds that society makes people with impairments disabled; that is, the social model makes a causal claim. However, some commentators are skeptical about the usefulness of the social model’s causal analysis (see Samaha 2007). Sean Aas and David Wasserman (2016) provide a forceful argument for why such a causal analysis is important: if society causes disability disadvantage, eliminating or mitigating such disadvantage has a higher priority vis-à-vis other disadvantages society did not cause. Their starting point is Nagel’s 1997 paper “Justice and Nature.” Nagel holds that society is not responsible for those disadvantages it did not cause. While Aas and Wasserman acknowledge that Nagel’s essay makes steps in the right direction regarding social responsibility, they rightly reject his radical narrowing of the purview of justice. They especially disagree with Nagel that society bears no responsibility for disadvantages if they are the unintended results of justifiable projects whose mitigation would be costly.

Aas and Wasserman lay down an alternative conception of social responsibility. They provide four criteria for when society is responsible for disadvantage. One of these states that society is responsible for a disadvantage if and only if the given disadvantage results from societal policies, which “in the course of making some better off, foreseeably make others, who have some claim to a justification of those policies, worse off.” Just like Nagel, Aas and Wasserman approach the question of just and unjust disadvantages from a tort law perspective—foreseeability as a responsibility-limiting factor is familiar from the tort law literature on “proximate cause.”  However, this literature makes it clear that using foreseeability as a responsibility-limiting factor can be problematic. Sometimes, agents are not responsible for disadvantages they can foresee and can be held responsible for causing disadvantages they did not foresee. Using the example of inaccessible public infrastructure, I argue that there are instances when the foreseeability criterion should be replaced with strict liability for causing unforeseeable disadvantages. In other instances, social actors are at fault for not foreseeing the caused disadvantages. As I aim to show in the paper, which of these cases applies depends on the available epistemic resources of agents in charge of providing access.