ABSTRACT

The standard semantics for subjunctive conditionals has been developed by David Lewis (1973a) and Robert Stalnaker (‘A Theory of Conditionals’ in Rescher 1968: 98-122). Roughly, the theory is as follows: we consider the nearest (metaphysically) possible worlds in which the antecedent is true and then we see whether the consequent is also true in those worlds. If it is then the conditional as a whole is true (in the actual world). What determines nearness to the actual world? It is partly a matter of past history, and partly a matter of having similar laws of nature. There is disagreement over whether there must be a unique closest world, or whether it is possible for there to be several worlds that are joint closest, or, indeed, whether it is possible that for every world there be a closer one. The most inclusive position we can take is to follow the lead of David Lewis and claim that a counterfactual conditional is true if and only if either the antecedent is false of logical necessity or there is a possible world in which both antecedent and consequent hold and that is closer to the actual world than any world in which the antecedent but not the consequent holds. On this view the proposition may be true in all three cases: if there is a unique closest world, if there are several equally close worlds, and if for every world there is a closer one.3 In what follows I shall just speak loosely of ‘the closest possible worlds’ to avoid circumlocutions. So in the examples in hand: what makes the proposition expressed by (4.1) true is that in the closest possible worlds in which the match has been struck it has lit. There are, of course, possible worlds in which the match has been struck but has not lit; it might have been prevented from lighting by a force 10 gale or it might have been a dud. These are remote possibilities in the situation envisaged. (They are not remote to other situations, of course; suppose that I had been given a dud, then the proposition expressed by (4.1) would have been false. Likewise, suppose that I were trying to light it in a severe gale then the proposition expressed by (4.1) would have been false. But these suppositions are in turn far from the actual situation, in which I have a box of good matches and a windless day.) Again, there might have been different laws of nature such that struck matches did not light, but this is a remote possibility. In other words, the possible worlds in which the actual laws of nature hold are, obviously, a lot closer to the actual world than those worlds in which the actual laws of nature do not hold.