But what we have now to consider are cases of wholes, in which one or more parts have a great negative value—are great positive evils. And first of all, we may take the strongest cases, like that of retributive punishment, in which we have a whole, exclusively composed of two great positive evils—wickedness and pain. Can such a whole ever be positively good on the whole? (§ 130 ¶ 1)

(1) I can see no reason to think that such wholes ever are positively good on the whole. But from the fact that they may, nevertheless, be less evils, than either of their parts taken singly, it follows that they have a characteristic which is most important for the correct decision of practical questions. It follows that, quite apart from consequences or any value which an evil may have as a mere means, it may, supposing one evil already exists, be worth while to create another, since, by the mere creation of this second, there may be constituted a whole less bad than if the original evil had been left to exist by itself. And similarly, with regard to all the wholes which I am about to consider, it must be remembered, that, even if they are not goods on the whole, yet, where an evil already exists, as in this world evils do exist, the existence of the other part of these wholes will constitute a thing desirable for its own sake—that is to say, not merely a means to future goods, but one of the ends which must be taken into account in estimating what that best possible state of things is, to which every right action must be a means. (§ 130 ¶ 2)