Intertwined Fates and Why Ex-Ante Contractualism Cannot Handle Them

Erik Ossenkopp

Saarland University (Universität des Saarlandes)

Let epistemic ex-ante contractualism (EEAC) be the claim – a variant of which is defended by Johann Frick – that, other things equal, an action is right iff it minimizes the sum of the strengths of the relevant complaints against it, where (i) the strength of a person’s complaint against an option is i.a. a function of her expected utility under that option, (ii) expected utilities are determined using epistemic probabilities, and (iii) roughly speaking, a complaint is relevant iff it is sufficiently strong. The literature on EEAC discusses cases of the following kind, where n is a natural number greater than 10 and m a real number between 0 and 1: »The agent has the choice between vaccinating n inhabitants against a disease and not vaccinating anyone. The vaccine gives each inhabitant a chance of surviving unscathed equal to 1-m but also a risk of death equal to m. Without the vaccine, all will be infected and lose a finger.« Advocates of EEAC like Frick argue that, if m is small enough, the agent ought to vaccinate, no matter how large n is, since only vaccinating maximizes each inhabitant’s expected utility and is thus in everyone’s best interest. This is a powerful argument. Yet, as the literature on EEAC points out, if n is large enough, the same line of thought justifies vaccinating in cases of the following kind: »It’s the same as the first case, except this time the vaccine kills anyone who has gene X and saves those who don’t. It is known that exactly ten inhabitants have gene X, but it is unknown who.« This time, many find EEAC’s verdict implausible, especially given that EEAC would deem vaccinating wrong if it were known who has gene X. Why does the same attractive line of thought lead to the wrong result? I argue that in the second case, contrary to the first one, there is a conflict of interest which EEAC leaves unaddressed. I offer a variant of EEAC that handles this issue by tackling probabilistic dependencies between people’s outcomes.

Chair: tba

Time:

Location: