Abstract

By Lewis's account, it may be unduly hard for subjunctive conditionals to be outright true. We could weaken them by deeming each one to be true if its antecedent makes its consequent highly probable. Or we could follow Edgington's proposal that subjunctive conditionals should not be thought of in terms of truth at all but only of probability. It is shown here that adopting this proposal would not significantly narrow the chasm between subjunctives and indicatives.