COUNTERFACTUALS AND CAUSALITY
D. S. Brée · Journal of Semantics · 1982
There are three parts to this paper. In the first part the difference between hypothetical and counterfactual conditionals is examined. Both Adams's argument that indicative and subjunctive conditionals differ in the degree to which they are justified and Lewis's contention that counterfactuals differ from hypotheticals in that they fail to contrapose are both shown to be unfounded. Standard tests confirm Karttunen's claim that the difference lies not in the truth conditions but in the falsity of the antecedent being presupposed and the truth value of the consequent being a conversational implicature. There is also a pragmatic difference: counterfactuals are more difficult to verify. In the second part 4 solutions to the counterfactual problem are criticized partly in light of the differences found in the first part. Goldstick's extension of Goodman's classic solution is shown to be an incomplete algorithim. Lewis's possible world solution reduces the problem to finding a criterion by which to order possible worlds. Krat-zer's solution, also in the possible worlds tradition, requires a heuristic for partitioning propositions. It is only the older solution of Rescher and Simon, based on causal ordering, that is adequate for causally based counterfactuals with false consequents. In the last part of the paper Simon and Rescher's method is extended to accidental counterfactuals and counterfactuals with true consequents.