As noted above, the Bennett algorithm cannot be applied to all counterfactuals. It requires that the antecedent take place at some particular time, and so fails to deal with, for example, the conditional If gravity had obeyed an inverse cube law, Kepler’s second law would still have held. It also requires that the antecedent be accessible by way of a conservative deviation, and so perhaps fails to handle, for example, If Caesar had directed the invasion of Iraq, there would have been no looting of the archeological museum (though Bennett argues that his full treatment applies to many conditionals of this sort). Finally, it does not deal with what Lewis calls backtracking counterfactuals, conditionals that implicitly or explicitly call for the ramifications of a counterfactual antecedent to be propagated into the past as well as into the future. Some of these cases — the first in particular — are of scientific interest (though the fact about Kepler’s second law and others like it can be expressed, I think, without recourse to counterfactual conditionals). But most scientifically significant counterfactuals, including the examples given earlier in this section, are amenable to the Bennett approach; it is, then, a solid foundation for a discussion of counterfactual support.
« Prev Next »
Top of page Top of page