Fortune Teller. When Vera is ten years old, Fortuna tells her that at some point in her life, Ivan, who will be born in five days, will pose a threat to her life and she will need to kill him to prevent him from killing her. Vera moves to a different town. Twenty-five years later, when traveling on vacation, she sees Ivan at a store. Two days later, Vera trips and falls into a well. Vera’s twin sister, Vena, who has always hated Vera, sees Ivan walk by the well and shoves him down the hole. Vera has, ever since her fortune was told, brought a ray gun with her so she can disintegrate Ivan and does so.
If Vera has a right to kill Ivan because he will otherwise violate her right not to be killed, that fact is true, not just once Vena pushes Ivan down the well, but two days prior, and even twenty-five years prior. It appears that Thomson’s view would justify Vera killing Ivan at birth (or perhaps in utero).58
Total Recall. Dena is driving a Loyota, a car brand that has recently been in the news for a failure of power steering. Victor is an investigative reporter who is about to publish an article on how all red Loyotas have this problem and the power steering fails two minutes after the driver pumps his brakes twice quickly. As he is walking through a tunnel, he sees Dena’s red Loyota and sees Dena stop quickly for a red light by pumping the brakes twice. In this tunnel, Victor is certain (and correct) that once Dena loses control of the car, she will come right toward the pedestrian walkway, swerve toward him, and, without intervention, kill him.
McMahan’s view is that once one chooses to engage in risky behavior, one is liable to defensive force if the activity turns out to be one that will impose harm. Now, McMahan may be able to prevent a total pushback along the time line, as he may be able to argue that until Dena chooses to drive the car, she is not morally responsible for the threat (as most criminal law theorists and moral philosophers create room for compatibilism) but her responsibility for the threat certainly obtains at the moment she gets in the car, as opposed to at the later moment for the brake failure. So, it would seem that Dena is morally liable to defensive force by Victor even prior to the power steering failure.
Hence, any theory of self-defense will have to have a theory of aggression that specifies how far the attacker must come. In asking this question, we must bear in mind that we are dealing with liability and not just permissibility. We are asking the conditions under which an aggressor gives up a right not to have deadly force used against him. This idea may (but does not necessarily) come apart from how much risk it is fair to ask the defender to bear.
Now, one possibility is that once the aggressor has formed a culpable intention, then the defender is permitted to act when she needs to. But “need” is a dangerous category. A defender may need to act prior to the formation of a culpable intention by the aggressor or the acquisition of any weapons necessary to launch an attack. If the right to self-defense is simply a right to act as early as is necessary to defend oneself effectively, then the need to defend may very well arise far before the initiation of any aggression.59
Indeed, consider how McMahan tortures his own view of liability to get to need-based cases. In Killing in War, McMahan imagines soldiers who join their country’s army for good reasons. Their government is plotting an unjust war against a neighboring country, but the soldiers are oblivious to this fact. May the neighboring country preemptively attack the soldiers? McMahan thinks so. “They earlier made a voluntary choice that in effect committed them in a public way to obedience, and those to whom they owe obedience will, unless prevented, order them to fight in an unjust war in which it is reasonable to expect they will participate.”60 I find this result troubling. The soldiers have not chosen to engage in any risk imposing activity, nor have they been given the opportunity to refuse to engage in an unjust war. I do not see why the soldiers have, at this point, forfeited their moral complaint.
One might think that the culpability view solves this problem because it requires need and culpability. Let us disentangle three different ideas. First, agents may have agent-relative permissions to prefer their lives to others’ when they need to. This idea does not require culpability. Second, in cases where one person needs to die, an individual’s culpability in creating the situation is one significant, though not the only significant, consideration.61 Indeed, it may be that as a matter of selection, life’s worth and overall culpability ought to matter.62 Third, if someone intends to kill you and you need to act at the moment of intention formation to stop him, you may act then. This third proposition is deceptively simple. As Larry Alexander and I have often argued, there are a number of problems with holding people criminally liable for early attempts – particularly an act like intention formation.63 It is difficult to distinguish intentions from fantasies; intentions are often (always?) internally and externally conditional; and intentions are revocable. Alex may intend “to kill Betty” but change his mind when he sees her smile (an externally conditional intention). Or Alex may intend “to kill Betty unless she smiles” (an internally conditional intention). Or Alex may intend to kill Betty, and then just simply think the better of it. Larry Alexander and I have argued that we cannot punish Alex based on the prediction that he will act because he very well may not.
Now, one way out of this dilemma for self-defense is to argue that it must be the case that Alex would have killed Betty for Alex to be liable. If Betty makes a mistaken prediction, she may not be punished (and legal theorists can debate whether a reasonable mistake makes her justified or excused), but Alex is not liable because he was not really going to do it. The thought is that “need” solves the dilemma. That is, we ignore all problems with “what if the aggressor would have changed her mind” by presupposing with “need” that she won’t. If we are entitled to assume necessity, then it appears that there should not be a requirement of action. That is, between an aggressor and a defender, if the defender must kill the aggressor, he is entitled to do so when the aggressor has formed what is, at that time, a culpable intention, and will remain a culpable intention in its execution. Between the innocent defender and the culpable (future) aggressor, the aggressor’s own choice is sufficient to alter his moral standing and allow defensive force to be used against him.
We might revisit our liability formulation thus:
When Awill purposefully, knowingly, or recklessly kill D, and A lacks justification or excuse, A is liable to defensive force once A forms a culpable intention and D must kill A to prevent being killed himself.64
Still, there is something altogether odd about premising the liability conditions of the aggressor on whether it will in fact be necessary to kill him – a fact that the defender cannot know, but can only predict. In the punishment context, this raises the question of pre-punishment.65 That is, we might worry about someone deserving something before his criminal act actually happens. In the context of self-defense, the concern is different. The concern is that a moral principle that relies on necessity ignores the very sort of practice that self-defense is. Self-defense can never be about whether the aggressor would have succeeded because self-defense is a moral permission to preempt an anticipated harm.
That is, I think that we can no more make sense of the question of whether we can pre-punish someone for a criminal act he has not done, then whether we can justify preempting a harm that definitely will occur. Both questions are confused. Punishment is responsive to an offense committed. Prevention is responsive to a threat perceived, and thus, for self-defense, the question should be what the offender must do that creates such an appearance such that he is liable to defensive force.
In other words, objective necessity is not a requirement of liability. Consider: