I must first clear up an ambiguity in the phrase ‘doing evil that good may come’. We cannot ask whether e.g. Caesar’s death was a good or bad thing to happen; there are various titles under which it may be called good or bad. One might very well say e.g. that a violent death was a bad thing to happen to a living organism but a good thing to happen to a man who claimed divine worship, and this would again leave it open whether doing Caesar to death was a good or bad thing to do for Brutus and the rest. Now when I speak of ‘not doing evil that good may come’, what I mean is that certain sorts of act are such bad things to do that they must never be done to secure any good or avoid any evil. For A to kill a man or cut off his arm is not necessarily a bad thing to do, though it is necessarily bad that such a thing should happen to a living organism. Only by a fallacy of equivocation can people argue that if you accept the principle of not doing evil that good may come, then you must be against capital punishment and surgical operations.

Suppose that A and B are agreed that adultery is a bad sort of behaviour, but that A accepts the principle of not doing evil that good may come, whereas B rejects it. Then in A’s moral deliberations adultery is simply out: as Aristotle said, there can be no deliberating when and how and with whom to commit it (EN 1107a16). For B, on the other hand, the prima facie objection to adultery is defeasible, and in some circumstances he may decide: Here and now adultery is the best thing. Similarly, Sir David Ross holds that the objection to punishing the innocent, viz. that then we are not ‘respecting the rights of those who have respected the rights of others’, is only a prima facie objection; in the general interest it may have to be overruled, ‘that the whole nation perish not’—a Scripture quotation that we may hope Sir David made without remembering who was speciously justifying whose judicial murder.

(Peter Geach, “The Moral Law and the Law of God,” chap. 5 in Absolutism and Its Consequentialist Critics, ed. Joram Graf Haber [Lanham, MD: Rowman & Littlefield Publishers, 1994], 63-72, at 65-6 [italics in original] [essay first published in 1969])

Note from KBJ: Person A is an absolutist deontologist. Person B is a moderate deontologist. They agree that adultery is intrinsically wrong (that’s what makes them deontologists rather than consequentialists), but disagree about whether it is ever justified, with A saying no and B saying yes. You might say that both A and B endorse a rule against adultery. For B, the rule has exceptions. For A, it does not.