My defense of Cornman survives Martin and Ruf's rejoinder. The basic point of my argument was neither that Martin and Ruf fail to provide a distinction between Jones and the other people in the boat, nor that the distinction they provide is insufficient to generate a utilitarian rule from which 'Jones ought to be sacrificed' can be derived. My basic point was rather that their distinction (and, indeed, any similarly objectionable distinction) is insufficient to generate a utilitarian rule which satisfies condition (2)(a) of Cornman' s principle.
A brief, and hopefully dearer, restatement of my argument is the following. Let a CL case be a case in which there are ten people in a lifeboat one of whom must be sacrificed if any of the ten are to be saved. The rule 'In a CL case, one of the people on the boat ought to be sacrificed' may be presumed to be a utilitarian rule conforming to Cornman's principle; but this rule, as I argued in my original defense of Cornman, is not strong enough to yield 'Jones ought to be sacrificed'. To derive 'Jones ought to be sacrificed' we need a rule of the form 'In a CL case, if exactly one person on the boat has property P, that person ought to be sacrificed'. But if a rule of this form is to be a utilitarian rule conforming to Cornman's principle, P must be instantiated in such a way that the resulting rule not only (A) is justified by utilitarian considerations, but also (B) does not violate condition (1) of Cornman's principle. And Martin and Ruf's argument fails precisely because the rule required by their alleged counterexample does not satisfy (B). Their example requires such a rule as:
(R1)
In a CL case, if exactly one person on the boat has received the lot marked 'victim" apparently as the result of a fair drawing but really on the basis of skin color, that person ought to be sacrificed.