Consider one of the most famous hypotheticals on the subject of self-defense: the Fat Man puzzle. In Fat Man, you find yourself in a small boat at the bottom of a chasm. Although there are many versions, what they have in common is that an enormously fat individual is hurtling down from the cliff. You have no idea why he is falling—whether, say, he jumped or was pushed. All you know for sure is that if he hits you, you die. You have no space to maneuver, and no time to escape. Fortunately, you are armed with your trusty Fat Man gun. You can pull the trigger and vaporize him, thereby saving yourself.
Theorists of self-defense usually posit that killing another to protect the self must be based either on the status of the attacker (e.g., enemy soldier in war) or what the attacker is doing (e.g., actively shooting at you). The Fat Man problem usefully divorces the justification for violent self-defense from the motive of the assailant. Robert Nozick’s original version of the problem stipulated that Fat Man has been pushed, and is therefore morally innocent; thus theories of self-defense that depend on what the attacker is doing (e.g., is he engaged in aggression?) cannot justify the use of the vaporizer.*
And yet the Fat Man problem is in other ways too easy. Augustine, to take an example, would surely have rejected the use of the vaporizer gun, on the ground that your life is not intrinsically more valuable than the Fat Man’s. Liberalism’s refusal to weigh lives against each other also makes calculation difficult. Yet I find that my students have little difficulty with the problem, answering as Nozick intended: they are by and large perfectly willing to blow Fat Man to smithereens to save themselves.This situation is exactly identical to a non-theoretical problem that we've discussed here just recently: the case of abortion where a zygote is wrongly implanted in the mother. Without needing the machinery of a "Fat Man Gun," we can see that the moral issues are fairly straightforward:
1) Party X is going to be killed by Party Y unless Party Y is killed first.
2) Party Y is doomed anyway.
This case brings out two issues that are important to ideas about the use of force. The first is that innocence is sometimes not a purely moral issue: sometimes innocence is practical. The baby is surely morally innocent, but practically the baby (or the Fat Man) is going to kill someone. Thus, we can consider using force against a kind of target that normally would be exempt from consideration.
The second is one that any ancient thinker would have understood: doom changes moral calculus. We don't have a case here where we can save Party Y. We are either allowing them to die, or killing them ourselves. The outcome for them is not different. Thus, we do our moral work on the issues that we can affect.
Taking action against such a target is a tragedy, but it may also be a duty.
* The footnote here says that "theories that rest on moral culpability would not justify shooting down an airliner carrying 100 innocent passengers and 3 hijackers, when the hijackers intend to fly into a building, killing everyone on board, and hundreds or thousands more on the ground. This is not to say that shooting the airline down cannot be justified; the calculus relies on a combination of consequentialist body-counting and double effect." But this is not correct: there is no need for a "combination" of this sort, because one of the components of the medieval doctrine of double effect is proportionality. Modern consequentialist thinking does not need to be added to make sense of the calculus; St. Thomas Aquinas established standards for dealing with that aspect of the question.