Proportionality in Self-Defense

Abstract

This article considers the proportionality requirement of the self-defense justification. It first lays bare the assumptions and the logic—and often illogic—underlying very strict accounts of the proportionality requirement. It argues that accounts that try to rule out lethal self-defense against threats to property or against threats of minor assault by an appeal to the supreme value of life have counter-intuitive implications and are untenable. Furthermore, it provides arguments demonstrating that there is not necessarily a right not to be killed in defense against theft or minor assaults. While there is a general moral right of self-defense and a general right to life, the scope of these rights (like the scope of the right to liberty and the scope of the right to property) depends on certain social facts that—even within a liberal framework—can differ from one society to another. Moreover, the proportionality of self-defense does not depend on the rights of the aggressor alone, but also on a precautionary rule, shaped by the balance of interests of the society in question and aimed at protecting innocent people and other social interests. This rule can protect an aggressor even in cases where he does not have the right to such protection.

Introduction

Some jurisdictions, at least in the past, had no official proportionality requirement for the self-defense justification at all. German law in the Weimar republic deemed it permissible (and the courts decided accordingly in at least one notorious case) to shoot at (and possibly kill) an apple thief if the defender (of the property) had no other means of stopping the thief. (van Rienen 2009: 181–182) A German legal adage applied to self-defense holds that “Right need not yield to wrong.”¹ This is, of course, the most lenient (towards the defender, and thus the harshest on the aggressor) view of proportionality one can have: it denies that there is a proportionality requirement in the first place. The strictest view, in contrast, is that the proportionality requirement does apply to self-defense but that self-defense is never proportionate. Yet this would not merely restrict the scope of the self-defense justification but instead abolish the self-defense-justification itself. It would not be a limiting condition on self-defense, but an “impossibility condition” of justified self-defense—it would make justified self-defense morally impossible.

Nobody (at least nobody I have ever encountered, either in literature or in real life) subscribes to the latter view, not even pacifists, who for the most part are not so much against any use of force but rather against war. Note, moreover, that even the claim that the use of force in self-defense is never proportionate would not yet completely rule out the justifiability of self-defense since not all self-defensive measures must avail themselves of the use of force. Consider, for example, a case where someone tries to hack into your bank account, and you stop him by simply turning off his power supply. This is an act of self-defense (namely of defense of your property or privacy), but it does not involve the use of force.

Yet even if one ignores one of the most extreme views on the proportionality condition as applied to self-defense because nobody holds that view anyway, the extreme view on the other end of the spectrum (namely the view that there is no proportionality requirement) remains a viable alternative that one cannot dismiss out of hand (at least not without thereby being simply dogmatic), and there are of course many other views to be encountered on this spectrum. For instance, while many of the Weimar scholars and judges deemed shooting a thief in defense of an apple justifiable, an Anglo-Saxon scholar, Fiona Leverick, writing more recently, declares in the context of a discussion of the proportionality condition that “the right to life cannot be forfeited by becoming a threat to property because no item of property is of such value that it outweighs the value of human life.” (Leverick 2006: 131) She then takes such a strict view on the proportionality condition that she not only rules out lethal measures in defense of property but subsequently struggles hard to make self-defense against rape justifiable: it is a struggle for her since it would seem that the value of human life also outweighs the value of sexual self-determination. Leverick, in turn, is aptly criticized by Uniacke (2011: 256–260), who points out that Leverick switches from one proportionality criterion to another, depending on what she wants to rule out or justify: she uses one criterion when she wants to rule out defense of property, and another one when she wants to justify defense against rape. Uniacke herself does not commit this mistake; she consistently applies one and the same proportionality criterion, according to which both “the harm and the corresponding wrong are relevant to a judgment about what would be proportionate self-defense.” (Uniacke 2011: 260) (The innocent victim is both harmed and wronged, but an aggressor is only harmed by just self-defense, not wronged, that is, his rights are not violated or infringed.) But then she claims that the “use of lethal force in fending off some types of unjust harm (e.g., minor assault) is clearly disproportionate given the seriousness of the harm inflicted on the attacker.” (Uniacke 2011: 262) However, this judgment does not follow from her account of proportionality, let alone “clearly” so. One commits no logical or conceptual error if one embraces Uniacke’s account and claims that lethal force against a minor assault is proportionate. Another author, Sangero (2006: 253), states that there “is no room in a civilised society for justification of saving mere property at the cost of human life,” yet fails to provide any real argument for it. (He remarks that property damages are reversible, but, first, that seems wrong if unique items are destroyed, like your Picasso, and small consolation if, in fact, they will not be reversed, whether in principle reversible or not—which is probably the statistically normal case for stolen items.) Likewise, Rodin (2002: 43) wants to “defend” the view “that there is a deep moral distinction between serious attacks against the person, on the one hand, and attacks against property rights, on the other, such that one may rightfully kill in the first case, but not in the second.” However, he at best tries to defend this view but then admits himself that “the considerations that [he] has brought forward are not so much reasons for why we should value life and property in a certain way, but rather pointers to the fact that we do value them in this way” and ends up with merely claiming that here “we have reached an ethical bedrock” that “we … discover by reflecting deeply on the case.” (Rodin 2002: 46) Yet while he might have reached his own ethical bedrock, we haven’t. There are, after all, quite a number of people who simply do not share Rodin’s intuitions. This includes certainly many German legal scholars (and probably many Russian ones) (Fletcher 1978: 871–872), but, according to my experience, also many “men and women on the street”—as opposed to Anglo-Saxon liberal philosophy professors in the seminar room. Simply declaring a particular and not necessarily widely shared intuition or judgment on proportionality as “ethical bedrock” is the opposite of “deep reflection.” We should try to do better.

I will try to do better in several steps. First, in Sect. 2, I lay bare the assumptions and the logic—and often illogic—underlying very strict accounts of the proportionality requirement. I will focus on Leverick, but my arguments in this section apply to all accounts that try to rule out lethal self-defense against threats to property or against threats of minor assault by an appeal to the supreme value of life. I will then, in Sect. 3, provide arguments demonstrating that there is not necessarily a right not to be killed in defense against theft or minor assaults. While I subscribe to the idea that there is a general moral right of self-defense and a general right to life, the scope of these rights (like the scope of the right to liberty and the scope of the right to property) depends on certain social facts that—even within a liberal framework—can differ from one society to another. In Sect. 3, I argue that the proportionality of self-defense does not depend on the rights of the aggressor alone, but also on a precautionary rule, shaped by the balance of interests of the society in question and aimed at protecting innocent people and other social interests. This rule can, in turn, itself create rights, even of the aggressor, but it operates on its own and can still protect an aggressor even in cases where he does not have the right to such protection.

Proportionality and the Value of Life

Let us go back to Leverick’s assertion that “the right to life cannot be forfeited by becoming a threat to property because no item of property is of such value that it outweighs the value of human life.” Leverick comes to this conclusion via the following argument:

The premise of this text is that all human beings, even those who commit or attempt to commit serious crimes, have a right to life. The reason we are permitted to kill someone who threatens our life is that the right to life of the aggressor is temporarily forfeited by virtue of becoming an unjust immediate threat to the life of another. Starting from this premise, it is difficult to see how lethal force could possibly be permissible against lesser threats than a threat to life. (Leverick 2006: 131)

The problem with this argument is that the conclusion simply does not follow from the premises. The premises that human beings have a right to life and that this right can be forfeited by virtue of becoming an unjust immediate threat to the life of another person do not imply that this right can be forfeited under this condition only. (It also implies, pace Leverick,² nothing about permissibility. This point need not concern us here, though.³) Leverick’s conclusion, far from resting on an argument, seems to be a mere stipulation. Thus, when she considers a contrary position according to which one can also forfeit one’s right to life on grounds of attacking someone’s property and risking one’s own life in the process, she simply informs the reader that “[s]uch an argument is comprehensively rejected here on the basis that human life is always worth more than property.” (Leverick 2006: 135) Yet even if human life were always worth more than property—this by itself does not imply that the right to life cannot be forfeited by posing an unjust immediate threat against something of lesser worth. Rather, one would need an additional premise, like the following one:

AP: An attacker can only forfeit a right to a thing x (like life, bodily integrity, property) by becoming an unjust immediate threat to something of equal or greater value than x.

However, while Leverick implicitly presupposes this additional premise (AP) as long as she discusses self-defense against threats to property, she provides no argument for it. In fact, she herself contradicts this premise once she gets to discussing self-defense against rape.⁴ For the moment, however, let us dwell a bit more on the case of property. How plausible is the premise as far as defense of property is concerned?

It is not plausible at all. Consider the following case: Anna knows perfectly well that Luisa’s signed baseball bat is worth 3000 Euros. Luisa hates Anna, and unprovoked she attempts to smash Anna’s laptop (value: only 1000 Euros). Anna is preparing her Amazon trip and has a machete next to her, which she uses to cut the baseball bat Luisa is swinging in half, rendering it worthless. Let us assume this act of defense satisfied the necessity condition under the circumstances. Was the defensive act disproportionate, and did Anna violate Luisa’s, the aggressor’s, property right to her baseball bat by defending her own significantly cheaper property against her? Leverick would have to say exactly that if she really accepted the premise in question; but to the best of my knowledge there is not one jurisdiction on the planet that would agree with her on this point.⁵ Moreover, the idea that the act was disproportionate is also morally entirely implausible. If Anna does not want her valuable baseball bat destroyed, she should not use it to attack other people’s property. “You knew that my baseball bat is more valuable than your laptop, so you should have let me destroy your laptop instead of defending it by destroying my baseball bat,” seems to be a rather silly and impudent thing to say. In fact, the entire premise seems to be extremely unattractive, since it has the implication that Luisa and others like her can immunize themselves from legitimate self-defense against their culpable aggressions by simply choosing expensive enough weapons and attacking in situations where the destruction of those very weapons would be the defender’s only or at least mildest means of defense of his own property. The premise leaves innocent people morally defenseless against inventive bullies.

Consider also the following example:

Kidnapping: Villain tries to kidnap innocent Victim. Victim knows (for whatever reasons) that if Villain succeeds, he will imprison her for three years in his well hidden lair (and then let her go) but not otherwise harm her (the imprisonment conditions equal those of Western industrialized countries, she will even have cell mates of the non-innocent kind, but those won’t harm her either).

On Leverick’s account, Victim is not allowed to kill Villain in self-defense, even if this is necessary to prevent the kidnapping. After all, human life seems to be more valuable than escaping 3 years imprisonment (one would rather go to prison for 3 years than die, for instance). However, this is counter-intuitive, and Western jurisdictions would allow lethal self-defense under these circumstances. But let us assume for the moment that Leverick is right that lethal self-defense is impermissible here and that, as her implicit assumption AP states, an attacker can only forfeit a right to a thing x (like life, bodily integrity, property) by becoming an unjust immediate threat to something of equal or greater value than x. Then it follows that Villain is allowed to kill people—including law enforcement officers—who try to kill him in order to keep him from kidnapping the innocent person. After all, if his right to life is not forfeited by his kidnapping attempt, then he still has it, and Leverick states: “The reason we are permitted to kill someone who threatens our life is that the right to life of the aggressor is temporarily forfeited by virtue of becoming an unjust immediate threat to the life of another.” (Leverick 2006: 131) The potentially lethal defenders would become precisely such unjust (that is: rights-violating) immediate threats to the life of the Villain, and so, on Leverick’s account, he can permissibly kill them—indeed, any number of them (since each of the potentially lethal attackers poses a threat to Villain’s life, which is still protected by his unforfeited right to life). This seems to be a perversion of justice. Maybe Leverick is willing to bite this bullet, but it is probably safe to assume that most people will not feel so inclined.

Moreover, Leverick’s problems do not end here. Consider the following example, a revised version of the previous one:

Sequential Kidnapping: Villain is very strong and likes to snatch innocent people and then imprison them, for the mere fun of it, for three years in his well-hidden lair. Nobody, not even he, can free them from the automated super-prison before the three years are over. One can (for whatever reasons, maybe he has a special protective suit) only stop him from doing this by killing him. He has engaged in this activity for two decades, taking 7,000 victims and planning to take thousands more.

On Leverick’s account, none of the victims is allowed to kill Villain in self-defense. After all, again, human life seems to be more valuable than escaping 3 years of prison. Note that Leverick cannot aggregate harms or rights-violations here. To repeat, once more, on her account the right to life of the aggressor is temporarily forfeited by virtue of becoming an unjust immediate threat to the life of another. The Villain of the above example, however, only poses an immediate threat to the person he is in the process of kidnapping right now, not to those other persons he will kidnap tomorrow or next year. (It is also not necessary to kill him now to save the others: one could still kill him tomorrow.) Thus, the fact that he kidnaps thousands of people cannot make him forfeit his right to life, as long as he does the kidnapping sequentially. But is an account of proportionality in self-defense that condemns us under the circumstances to letting the kidnapper continue in his enterprise forever, and that allows him to kill in self-defense any number of people who try to stop him by lethal means, something we can reasonably accept? It would not seem so.

One problem with Leverick’s account is precisely that she does not aggregate harms or rights-violations. However, once one starts discussing “value” one should recognize that value can indeed aggregate and accumulate. If you have five one-dollar bills on the left side of your table and one one-dollar bill on the right side, then the value accumulated on the left side is five dollars, even if there should be a centimeter distance between each of the bills, and thus there is a bigger amount of money—greater value—on the left side than on the right. To apply this to the case at hand: if one endorses a proportionality criterion for self-defense that rules out practically effective self-defense in the Sequential Kidnapping case, then one is not only endorsing the view that the culpable kidnapper’s life is more valuable than the freedom of one of his innocent victims but also that the culpable kidnapper’s life is more valuable than the freedom of thousands of innocent victims. Yet such a value judgment is entirely implausible.

To begin with, it is not even clear that the kidnapper’s life is more valuable than the freedom of only one of his victims. While Leverick (2006: 81; see also vii, 2, and 46) states that it is “the fundamental argument of [her] book … that all human life is of equal value,” she provides no argumentative support for this fundamental premise of hers.⁶ This is not surprising, for on further reflection the premise is anything but convincing. To wit, imagine you are with your dinghy on the ocean and some cosmic fluke throws Nelson Mandela into your space-time continuum to the North of your boat and Stalin and Hitler to the South. You can only safe either Mandela or Hitler and Stalin (and you know that none of them will pose threats to others in the future). Whom would you save? If you are of the opinion that indeed all human life has equal moral worth, then you must, if you are impartial and on pain of inconsistency, save Hitler and Stalin and let Mandela drown. I find this utterly implausible, and in my experience pretty much everybody apart from a few egalitarian political philosophers agrees. This agreement, incidentally, is not affected if we modify the example to replace Nelson Mandela with an innocent rape victim and Hitler and Stalin with her two rapists who also happen to be murderers.

This modification brings us to another point: even if all lives had equal moral worth, this would actually not oblige one to be impartial with regard to those lives. A mother, for example, is justified in saving her own child instead of two unknown children if she cannot do both. She is also, for that matter, permitted to save herself instead of two strangers. Moreover, in most Western jurisdictions persons are at best required to take upon themselves very minimal risks and costs if it comes to saving others.⁷ But then why, exactly, should an innocent person be required to sacrifice her own freedom in order to spare the life of a culpable aggressor intent on stripping her of her freedom?

At this point, an objection might be that the example glosses over the difference between killing and letting die. In the dinghy example, the savior of Mandela lets Hitler and Stalin passively die; he does not save them—but he does not actively kill them. Many philosophers think that this distinction between killing and letting die is of great normative significance. (Howard-Snyder 2011) Thus, someone who shares the intuition that one may save Mandela instead of Hitler and Stalin could claim that this has little bearing on the case of self-defense, since Hitler and Stalin do not have a right to be saved while the aggressors in the Kidnapping examples do have a right not to be killed.

However, whether the aggressors do have a right not to be killed in defense of 3 years of liberty is precisely the contentious issue. Moreover, the interesting question with regard to Hitler and Stalin is why they do not have a right to be saved (and I agree that they do not). Consider, as a foil, some different cases. You are again in your dinghy, and a lone innocent person is drowning a meter away from you (no other people are nearby). You could easily pull her aboard, but you cannot be bothered and just watch her drown. Here it seems that you do violate a claim right of the innocent person, one that, in some jurisdictions, would be enforceable in both criminal and civil court.⁸ And apart from the most unflinching libertarians, most people would probably agree that the drowning person has at least a valid moral complaint against you if you just let her die in such a situation although you could save her at little cost to yourself. Yet the picture changes again if we add a second drowning innocent person to this example and you again can only save one of them. Does the first stranger still have a right to be saved by you? It does not seem so. The explanation that German law gives for this comes under the heading of “collision of duties” (Pflichtenkollision): since you cannot discharge both your duty to save the first stranger and your duty to save the second stranger, and law must not confront citizens with incompatible demands, you are released from one of these duties (if you decide to save the second stranger, the duty to the first one vanishes).⁹ Thus, while the dependents of the lone stranger could file a wrongful death suit against you (in some jurisdictions) if you let him drown in the first case, they cannot do this in the second case (at least not legitimately and with any chances of success). What, however, if there is one innocent person drowning on the left and two on the right? Interestingly, German law, according to majority opinion among German legal commentators, refuses to weigh lives against lives. Thus, under German law (setting aside a rather revealing qualification¹⁰), you would be free to save just one person even if the alternative would be to let ten other persons drown. A vocal minority among German legal commentators is rather displeased with this position, and so is a large part of the German population, as could be seen in the discussion about shooting down a hijacked passenger plane (thus killing the innocent people on board) to prevent it being used as a weapon of mass destruction.¹¹ Other jurisdictions, in contrast, accept the possibility of weighing lives against lives without qualms, and in many jurisdictions, including in the United States, a life guard who opted, all else being equal, for saving the smaller number of drowning people instead the larger number would probably also face wrongful death suits and criminal charges. This seems to be a position preferable to the one taken by German law, for how could numbers not count?¹² Empirical studies on the so-called Trolley Problem show that the overwhelming majority of people think that they do count. (Edmonds 2014: esp. ch. 9; Bruers and Braeckman 2014)

At this point we are able to return to our original Mandela vs. Hitler and Stalin case. My argument was, first, that this case shows that human lives do not all have equal worth: Mandela’s life is at least two times more valuable than Hitler’s or Stalin’s. My second argument—and this is what we are discussing at the moment—was that even if all (human) lives had equal worth or value, people are often nevertheless permitted to care more for some lives than for others (the mother is permitted to save her own child instead of two other children of equal value) and I suggested that therefore it is mysterious why an innocent person should be required to sacrifice her own freedom in order to spare the life of a culpable aggressor intent on stripping her of her freedom. In reply to this I had the opponent muse that it might be permissible to save something of lesser value (namely, for instance, merely one life) instead of saving something of higher value (two lives) if—and only if—by doing so one does not violate any rights. My response, finally, is the following: It seems that in a case where someone can at little cost to himself rescue either a larger number of people or a smaller number of people, the members of the larger group (all else being equal) do have a claim-right to be rescued—unless the members of the larger group are guilty of serious rights-violations (or there are members in the smaller group to whom the rescuer has special responsibilities). If they are guilty of such violations, then—especially given that one has to choose between their lives and innocent lives—they forfeit their claim-right to be saved. But then the above-mentioned mystery returns—in a slightly different form, but no less mysterious: if two people, on grounds of culpability,¹³ can forfeit their rights to be saved in a situation where the choice is between their jointly more valuable two lives and an innocent single life, why then can a culpable aggressor, in a situation where the choice is between his more valuable life and a less valuable good like an innocent person’s liberty, not likewise forfeit his right not to be killed? It appears to me that it will be exceedingly difficult to answer this question short of question-begging and ad hoc stipulations.

Of course, again, if someone oddly rejects the intuition that it is morally justifiable—I do not even need to claim that it is obligatory—to save Mandela instead of Hitler and Stalin from drowning, then the above line of reasoning will probably leave this person unmoved. Conversely, however, Leverick has also most certainly not adduced any argument that could move those who do share said intuition—nor has anybody else.

So far I have argued a) that it is implausible to assume that all human lives have equal value, and b) that it might be permissible to care more for some lives than for others even if they did all have equal value. This point (b) thus provides a different argument for a conclusion already reached above with the help of the baseball bat/laptop example: it is permissible to defend something of lesser value to the detriment of something of a higher value. This is no less true for properties and belongings of different values than for liberties and lives.

This latter point can now be further strengthened with an argument already adumbrated above in the context of the Sequential Kidnapping case. I stated that if one endorses a proportionality criterion for self-defense that rules out practically effective self-defense in the Sequential Kidnapping case, this not only means endorsing the view that the culpable kidnapper’s life is more valuable than the freedom of one of his innocent victims, but also that the culpable kidnapper’s life is more valuable than the freedom of thousands of innocent victims. And I contended that such a value judgment is entirely implausible.

Its implausibility can also be seen if we attend to Leverick’s discussion of the alleged “duty to retreat,” according to which a (potential) defender has to retreat (instead of using force) if he can do so safely. In support of her own endorsement of this idea, Leverick cites from a decision made by the Court of Appeals of Alabama in Cooke v. State.

It is better that one man should flee rather than take human life, and at a time like the present, when human life is taken upon the least provocation, and observance of the law is held so lightly, and the very fabric of our laws seem to be trembling in the balance, it is well that we should restate and emphasise the rule, that it is better and right that a man should flee rather than he take human life, if he can do so without apparently increasing his danger to life and limb. It is no cowardly doctrine. It is the law of his sovereign state, which he is bound to obey; it is the law of God, which he ignores at his peril. It is the reasonable law of God and man. (Leverick 2006: 78)

Leverick takes this quote as endorsing the supreme value of human life and commends the court for making its point “forcefully.” Alternatively, however, one could also chastise the court for engaging in hypocritical and sanctimonious cant. I say this because I severely doubt that the members of an all male court in Alabama in 1921 really valued human life so high as to appreciate, for instance, a law enforcement officer running away from a burglar trying to break into the house of one of the judges because the officer is concerned the confrontation might escalate into the use of lethal force if he stands his ground in defense of property and the law—something he, incidentally, is actually legally expected to do. I also doubt that they would appreciate American soldiers running away from their enemy in order to spare even their own lives—let alone the lives of the enemy. In fact, if one wants to keep “the very fabric of our laws” from “trembling,” one should not encourage law enforcement officers to retreat from criminals in order to safeguard the criminals’ lives at the expense of law and order. Thus, for all its appeals to “God,” the court is probably less expressing its valuation of human life here than its rather mundane valuation of the “sovereign state”: private citizens must retreat if that is what the state’s laws demand.

However, states usually do not really impose a full-fledged duty to retreat even on private citizens. (Card 2010: 711; Sangero 2006: 199–200; Liptak 2006) Moreover, not one Western jurisdiction demands that law enforcement officers retreat (and I doubt that non-Western jurisdictions do). (Sangero 2006: 202–203) Leverick never mentions this rather revealing fact. What it reveals, of course, is that jurisdictions might value law and order more highly than the life of the criminals who threaten this order. She does, however, briefly discuss the German dual theory of self-defense, according to which acts of self-defense defend not only the victim of an aggression, but also the social-legal order in general. German law commentaries also use this to explain the leniency of the German proportionality or, more precisely, no gross disproportionality requirement. Leverick, somewhat strangely, deems this a “consequentialist approach” (the approach is actually downright Kantian), and dismisses it with the statement that “consequentialist justifications of the permissibility of self-defensive force were rejected [earlier in her book] because they fail to pay adequate attention to issues of rights. (Leverick 2006: 135) Yet what is at stake here is not the justification of the very permissibility of self-defense as such, but rather the scope and limits of one of its elements, namely of the proportionality requirement. My own justification of self-defense is most certainly not consequentialist, but rather profoundly anti-consequentialist. (Steinhoff 2016a) That being stated, one certainly need not be a consequentialist to admit, and indeed emphasize, that consequences count in one way or the other.

To wit, one can evaluate at the very least the moral invalidity of imperatives by paying attention to what would happen if they were actually being followed. This is not the same with mere permissions. Often people use the “What if everyone would do it?”—question as an argument against a permission. But, of course, it is a bad argument. Just because terrible things would happen (people would squeeze each other to death) if everyone went every day to the Museum of Modern Art in New York does not mean that Bill Smith and Jenny Jones may not go into the MOMA every day. In other words, “What if everyone would do that?” is not a good objection to a moral statement that does not require everyone to do it—and permissions do not. They only allow people to do it. (It would be a much better objection, though, if all people were inclined to do it. But that is an additional premise; “What if everyone would do that” is inconclusive merely on its own.) Yet an imperative of the form “You ought to go everyday to the MOMA,” if addressed to everyone (like “You shall not covet your neighbor’s wife”), does require everyone to abide by it. Likewise, a proportionality rule requires everyone within its jurisdiction to abide by it. But a proportionality rule that enjoins people from ever killing in defense against an attack that is directed at something of less value than life does lead to terrible consequences. In the Sequential Kidnapping case it leads to leaving thousands of people defenseless against a culpable attacker who deprives each of them of their freedom for 3 years.

Note that this is not only so in the Sequential Kidnapping case. The threat might not come from one Sequential Kidnapper kidnapping many times but from many kidnappers kidnapping one time each. Abiding by Leverick’s rule would again lead to many kidnappings. In contrast, the social acceptance of a lenient proportionality requirement allowing the shooting of the kidnapper would have a deterrent effect, thus sparing thousands of people from the fate of being kidnapped while simultaneously, through deterrence, sparing thousands (but not all) kidnappers from being killed. The latter arrangement appears to be the preferable one.

Note that the argument presented here is immune to claims like this one, made by Rodin: “An important function of proportionality … is to place limits on the degree of punishment [the context makes clear that Rodin means to apply this statement also to self-defense] a person may properly be subject to whatever the attendant deterrence effects may be of punishing more severely. This implies that deterrence effects must be largely excluded from judgments of proportionality.” (Rodin 2004: 94) My point does not concern the deterrent effect of defending oneself against (or punishing) “a person,” that is, a particular individual; instead, it concerns the deterrence effects of a certain proportionality rule and of people’s using the leeway it gives. And the claim that deterrence effects must be “largely excluded” from judgments about the validity of a rule regarding proportionality in punishment or self-defense would be nothing less than an invitation to blatant irrationality.

Likewise, Rodin’s argument against aggregation is a red herring. He insinuates that allowing aggregation might have as a consequence the permissibility of defensively killing someone who is about to sabotage the global transmission of Baywatch and to deprive “a global audience of hundreds of millions … of the pleasure of watching Pamela Anderson and David Hasselhoff.” (Rodin 2004: 94) Yet balancing interests and aggregation need not be a simple-minded procedure. There could, for example (and there are other possibilities), be certain thresholds below which aggregation is not allowed. Thus the view that the life of the saboteur outweighs even the pleasure that hundreds of millions derive from watching Baywatch is quite compatible with the view that the years-long false imprisonment of thousands of people outweighs the value of the life of the aggressor who falsely imprisoned them.

Thus, even if one accepted Leverick’s (2006: 79) apodictic statement that the “life of even the racist attacker with genocidal intentions is always worth more than the pride, dignity, honour, or freedom of movement of another individual,” this would still be a far cry from accepting that it is also worth more than the pride, dignity, honor, freedom, or, for that matter, property of thousands of individuals. To drive this point home with yet another example: Imagine that this racist attacker constructs himself an Iron Man suit,¹⁴ which means that the only way to stop him is to attack him with a missile that would blow him into pieces, thus killing him. Imagine further that he announces his plan, and then acts on it, to destroy all black churches and all Jewish synagogues in the United States. Leverick would literally leave us morally defenseless against this racist: we would have to idly stand by while he enacts his vile plan—after all, he “only” destroys property (although a lot of it). And once he is through with destroying black churches and synagogues, he could continue his destructive spree by reducing modern art museums to rubble, without anybody able to permissibly stop him. Etc. While Leverick might perhaps see such defenselessness as the price to pay for “civilized society” (Leverick 2006: 79), it is doubtful whether any civilized society has ever shared this assessment. Some certainly do not.

Perhaps at this point one might muse that there could be another justification (that is: not the self-defense justification) to avoid the unpalatable result that one has to refrain from lethal measures against the otherwise unstoppable evil Iron Man of our example. But what justification could that be? The problem the Racist Iron Man case and the Sequential Kidnapping case raise for Leverick’s account and accounts like hers is that such accounts allow lethal defense only if it is necessary to protect something of as much value as life. As noted before, the fact that the evil Iron Man and the sequential kidnapper will in the course of their criminal career destroy (as the examples suggest) something of greater value than the criminal’s life cannot overcome this problem. To illustrate this, let us assume that 30 synagogues have exactly the value of one human life. Thus, killing Iron Man when he is destroying the fifth synagogue is not necessary to keep him from destroying something of greater value than life. This disqualifies the self-defense justification. It disqualifies the necessity or justifying emergency justification for the same reason: there is absolutely no need to kill him now to keep him from destroying 30 synagogues—after all, there are still 25 synagogues to go. Moreover, the situation does not change once the racist approaches the thirtieth synagogue. He has already destroyed the other 29 synagogues, so by killing him now one does not protect something that has the value of 30 synagogues and thus of life—the other 29 do not exist anymore and therefore the time to protect them has passed—one protects at best the one synagogue at stake now and the ones he would destroy in the future. However, in order to save those other synagogues from future destruction one need not kill him now—hence, neither the self-defense justification nor the necessity justification (given Leverick’s assumptions) apply. Therefore, an alternative justification for avoiding the unpalatable implication that one must allow racist Iron Man to destroy all black churches and Jewish synagogues in the USA (and elsewhere, for that matter) would have to drop the assumption that it is only permissible to kill someone if “the right to life of the aggressor is temporarily forfeited by virtue of becoming an unjust immediate threat to the life of another.” (Leverick 2006: 131) Instead, one could assume that an aggressor also forfeits his right to life if he has aggregatively destroyed something of equal value to human life. In that case (and if forfeiture implied permissibility, as it does in Leverick’s framework) one could kill the aggressor after he had destroyed the thirtieth church. One would then, however, also be permitted to kill him whether or not he continues on his destructive spree. One could kill him for merely trying to steal an apple, for instance. In fact, after the thirtieth destroyed church one could permissibly kill him for the mere fun of it while after the 29th church it was still not permissible to kill him in order to keep him from destroying the next church. This alternative justification just considered seems to have sufficiently bizarre implications to reject it.

Another alternative is to state that someone forfeits his right to life and can be permissibly killed after a time t₁ if he will aggregatively destroy something of equal value to human life during his life after t₁ unless one kills him at some point after t₁. On that account, however, one could permissibly kill racist Iron Man by drowning him already as a baby during his baptism. This, too, seems to be a rather unpalatable implication.

There might, of course, be other possible alternatives, but I doubt that they will produce acceptable results. More importantly, however: if one can only get to such alternatives by dropping the assumption that one can only forfeit one’s right to life by becoming an unjust immediate threat to something as valuable as life, then one can already drop this assumption in the context of the self-defense justification itself. In other words, instead of groping for new justifications that can take up the slack behind Leverick’s account of self-defense (a slack consisting of leaving people defenseless against wrongdoers who threaten things less valuable than life and can only be stopped by being killed), one should try to formulate an account of self-defense that does not produce such slack in the first place. We need an account that does not value the lives of a few aggressors higher than the rights of the vast innocent majority to liberty, property, and bodily integrity.

Proportionality and the Rights of the Aggressor

David Rodin claims [without providing any argument in support of this claim other than a reference to the authority of Jeff McMahan, who in turn only makes a stipulation (see McMahan 2009: 10)] that there “is no independent condition of liability to harm separate from considerations of proportionality, and no harm can be described as proportionate without reference to the liability of the person affected. This is because a person can only be liable to a particular harm that is proportionate in the circumstances (if the harm were not proportionate, he would not be liable to it).” (Rodin 2011: 79) I argue below that an aggressor can be liable to disproportionate counter-measures, namely to counter-measures that are disproportionate measured against the standard of proportionality as a precautionary rule.¹⁵ Of course, I also concede that an aggressor might have a right (that depends, as we will see) against certain counter-measures on account of their being out of proportion with the severity of his attack. Thus, on my account counter-measures can be disproportionate—on grounds of the precautionary rule—although they do not violate the aggressor’s right, while other counter-measures could be disproportionate on the basis of the aggressor’s rights.

Authors like McMahan, Rodin, Quong (2015: 144),¹⁶ and other rights-forfeiture theorists,¹⁷ in contrast, can only take this second route if they want to prohibit some defensive measure as disproportionate under the circumstances. In other words, if they think that killing someone in order to defend one’s finger or property is disproportionate, then these authors must thereby contend that a culpable aggressor attempting to break an innocent person’s finger or assaulting her to get her money has a right not to be killed. Unfortunately, as I have already pointed out above, we never get offered any argument or explanation as to why we should accept such a contention. All we get are assurances that that is part of our “ethical bedrock” (Rodin), concessions (that’s honest, at least) that competing theories are simply “comprehensively rejected” (Leverick), or sanguine affirmations that this is all “clear” (Quong and Uniacke, among others). In the following I would like to burst this bubble and provide some considerations as to why this actually isn’t clear at all.

It already becomes much less clear if we flesh out the notorious apple thief example a bit, as does Robert F. Schopp:

Suppose V is an elderly farmer who grows apples in an orchard and sells them at the local market. A is the local bully who walks up to the farmer’s stall at the market and takes the apple. When the farmer says, “Stop, that is my apple,” the bully responds, “Tough, old lady, what are you going to do about it?” The farmer, who is old, frail, and less than arm’s length away from the bully takes a gun from under her apron and shoots him. Does this victim shooting this apple thief provide the intuitively clear case of an immoral act that we usually accept as an intuitive counterexample? When completed with details that render A fully culpable and the shooting actually necessary to stop the culpable violation of V’s sovereignty without submitting V to additional risk, it does not seem counterintuitive to say that V was justified. (Schopp 1998: 84)

For the time being, however, we do not need to decide whether it was justified ¹⁸; rather, what we have to consider is whether the thief had the right not to be killed under the circumstances. And in doing so we should not just dogmatically insist on him having this right and claim that this is “clear,” but, for a change, have a look at considerations that speak against him having this right.

One such consideration is reciprocity. Given the character of A, it is likely that he would have shot V if the tables had been turned: if he could only have stopped V from stealing his, A’s, apple by killing V. Assume that this is really true, that this is indeed how he would have acted. How can he then have any “standing to complain” [to employ a phrase often used by Quong (2015: 146, 162, 163, 167)]? How can he have any standing to complain if V does to him what he, under the same circumstances, would have been only all too willing to do to V? It would seem he doesn’t. If, however, he has no standing to complain about V killing him, then, it would seem, A does not have a moral right that V does not kill him [and Quong (2015: 146, 162, 163, 167) would agree].

The invocation of reciprocity should also appeal to Rodin, who, after all, defends a “reciprocity theory of rights.” (Rodin 2014) He claims that “I have the right that you not kill me because and to the extent that I comply with your right that I not kill you.” (Rodin 2014: 287) While the positive part of this theory (the part explaining rights-possession, expressed by the “because”) is clearly wrong since it is obviously circular (where is your right coming from—from mine?) (Fabre 2014: 397–398), the negative part of this theory (explaining rights-forfeiture) is correct and a stock in trade of the natural law tradition and German Idealism.

Furthermore, while the claim just quoted refers to actual compliance, Rodin is also, especially in the context of assistance, interested in “counterfactual reciprocal compliance”: “if it is counterfactually true of a person in need, X, that he would have assisted Y were their circumstances reversed, then this can generate an obligation for Y to provide comparable assistance to X.” (Rodin 2014: 296) And he then asks a question that is obviously also highly pertinent in the present context: “[S]uppose I am armed and I see my enemy walking unarmed down the street. He is currently complying with my right to life, but I have very good reason to believe that were our circumstances reversed (were he armed and I unarmed) he would shoot me down. Does the counterfactual reciprocity test provide any justification for my shooting him?” (Rodin 2014: 299) His answer is: “Clearly it does not.” (“Clearly” again!) His reason: “Actual compliance always trumps counterfactual compliance.” (Rodin 2014: 300)

Why is that? Because there is allegedly an “ontological” and “epistemological” difference between the two forms of reciprocity, that is, between “counterfactual reciprocity of compliance” and “actual reciprocal compliance. (Rodin 2014: 302–303) As regards the “ontological” difference, Rodin states: “Most people we are called on to assist are not currently assisting us and nor have they assisted us in the past. Instead, the obligation to assist them arises from the counterfactual claim that they would assist us in comparable circumstances.” (Rodin 2014: 303) But this is wrong: the obligation does not arise from a “claim.” It does not matter what claims people make. Rather, the obligation arises from what others would do for us; that is, they arise, for example, from it being “true of the poor man now that he would assist were he to be rich, and the rich man was poor.” (Rodin 2014: 299) In other words, it arises from another person’s dispositions to act, as Rodin himself acknowledges. (Rodin 2014: 298) And while dispositions to act are not acts, they are also not “counter-factual” but factual: people actually have them now. Thus, the alleged “ontological difference” is not particularly striking.

As regards the “epistemological” difference, Rodin (2014: 303) states that “I can know with near certainty that you are currently complying with my right to life,” while “the counterfactual claim about your free actions undertaken under alternative circumstances can only be known with a much weaker degree of certainty.” However, first, that is often simply wrong (people in the cross-hairs of an assassin with a sniper rife will normally not have any knowledge that they are under attack) and, secondly, irrelevant: that the witness does not know that the mobster assassin is about to kill her does not change anything about the fact that the mobster is liable to attack, even lethal attack if necessary. If the witness played around with her own gun and shot the assassin just as he is about to pull the trigger, the assassin would hardly be in a position to complain.

In the light of these considerations it is difficult to see how the apple thief, if he, in turn, would kill V if she tried to steal an apple of his, can possibly be in a position to complain. If he does, and with his last breath cries out: “You terrible person, how could you …?”, the appropriate and quite deserved answer is: “Look who’s talking. You would without compunction have done exactly the same. The only difference is that I would not have tried to steal your apple in the first place. But if I had, you would have killed me. So who are you to complain?”

Maybe Rodin, or someone on his behalf, might be inclined to object that the apple thief does have a valid complaint here: after all, he did abide by his duty not to kill V, even while stealing the apple, and according to Rodin’s “reciprocity theory of rights,” this is what gives rise to V’s duty not to kill the thief. Yet as already pointed out, Rodin’s theory is mistaken. To wit, even Rodin himself (Rodin 2011: 105) admits that apple thieves and other “lesser aggressors” are liable to some defensive measures. It would appear, for example, that the thief is at least liable to the victim’s constraining his movement by grabbing his arm. Yet the apple thief himself is only stealing an apple, and while doing so he abides by his duty not to constrain V’s freedom of movement: he is not grabbing V’s arm. But if a culpable aggressor, even in the relatively harmless form of an apple thief, forfeits his right not to be constrained although he abides by his duty not to constrain and thus disproves the positive part of Rodin’s theory, then that same theory cannot be used to counter the claim that the apple thief of our example is liable to lethal force even if he does not employ lethal force himself. Thus, considerations of reciprocity speak for the idea that the apple thief of our example has forfeited his right not to be killed in necessary defense of property, not against it.

What, however, if the apple thief would not have shot if the roles had been reversed? Would he then have a complaint against V? Well, there seem to be still other ways one can forfeit one’s rights or at the very least severely reduce their stringency. Imagine, for example, that someone walks around propagating the idea: “We should really hang every horse thief we get.” Even if she would never act on her own advice, it would appear that she has very poor standing to complain if she gets hanged after she has stolen a horse. The same is true for someone propagating killing thieves in self-defense. (Just to avoid cheap ad hominem arguments: so far I have not propagated the idea of killing thieves. Instead, I am arguing that killing thieves is not necessarily a violation of their rights.) This might, moreover, even be true for someone who simply thinks that the practice of killing thieves is unobjectionable. After all, if she thinks that it is unobjectionable but complains anyway if she is about to be killed for her thievery, she seems simply to be a hypocrite.

But suppose that someone thinks that killing thieves is wrong, would not do it herself, and she does not propagate the idea of doing it. Would she be wronged if one killed her in necessary defense if she tries to steal something? The answer is: it depends. If she is stealing in a society in which it is deemed acceptable to kill in defense of property, she has, I submit, no standing to complain and thus her rights are not violated—provided the society is reasonably fair and the thief does not belong to a group that has been systematically and unjustly been disenfranchised by the reigning property relations. If, in contrast, she steals in a society in which it is not deemed acceptable to kill in defense of property, her rights would have been violated.

Before explaining why this is the case, let me first say something in defense against the indignant outcry that the last paragraph is likely to evoke: “That’s relativistic!” Of course it is “relativistic” in a certain sense, but this relativism is of a rather unexciting and widely accepted run-of-the-mill type. To wit, most people seem to be of the opinion that collecting 7% VAT on book purchases is just (that is, it does not involve any rights-violation) if a legitimate government has decided to collect 7%. They also seem to think that collecting 14% VAT is just if the government has decided to collect 14% instead. Indeed, it is a fact that different governments raise different taxes on these kinds of purchases, and few people seem to deem this unjust—they do not think that there is some absolutely right percentage as regards VAT for book purchases and the like. Thus it would appear that the justice of certain things is determined by “the laws of the land” and not by some abstract standard hovering above the land. The proportionality requirement of self-defense, I submit, is one of those things.¹⁹

Interestingly—and perhaps inconsistently, given how absolutist she is or at least appears to be with regard to “the value of life”—Leverick thinks the same. I already pointed out that she has severe difficulties combining her endorsement of lethal self-defense against rape with her explicit equivalent harm view of the self-defense justification. Being raped, after all, does not seem to be a harm equivalent to being killed. Being killed is worse. This is why rapists can coerce their victims into compliance with credible death threats, and it is also why murder is punished more severely than rape. Leverick’s (2006: 157) explanation as to why “law should provide a complete defence to the accused who has killed to prevent rape” is that “rape approaches the standard of a wrong equivalent to a deprivation of life itself.” (Leverick 2006: 157)²⁰ This, of course, exemplifies precisely the shift Uniacke complains about in her critique of Leverick: while before Leverick constantly talks about “harm” and “value,” now, in order to justify lethal self-defense against rape, she suddenly talks about “a wrong.” Yet that does not even solve her problem.²¹ For it is simply not true that “rape approaches the standard of a wrong equivalent to a deprivation of life itself.” Again: if it did, why should people prefer being raped to being killed, and why should law punish murder more severely than rape?

One could, of course, just appeal to some metaphysical fact of the matter: “That’s just how it is, rape is as bad as murder, period,” but such dogmatic stipulation will hardly be able to convince anyone who is not (irrationally) convinced already. Fortunately, Leverick herself does not appeal to any absolute metaphysical truth, but rather to “the significance with which society has imbued the act of sexual penetration.” (Leverick 2006: 157) In other words, according to her, rape is morally as bad as murder in a society because society sees it that way. Yet it is obvious that at least our Western societies do not “imbue” sexual penetration and thus rape with such significance, which, again, is shown by the fact that our societies do not punish rape as severely as murder and by the fact that most people (even in other societies) would prefer being raped to being killed. Be that as it may, the important point in the present context is that Leverick (2006: 157, n. 94) clearly recognizes the social relativist implications of her view: “If the wrong is socially constructed, this does raise the question of whether rape would be equally wrong in a society where sexual penetration was not imbued with such significance.”

Indeed, it does. However, if the wrong of rape can be “socially constructed,” so can be the wrong of theft. Thus, if a society imbues the taking of another person’s property with such significance as to permit the killing of the thief, then taking another person’s property has this significance in that society and therefore killing the thief would, in that society, be permissible on account of Leverick’s very own social construction theory of the gravity of (some) moral wrongs.

I said “social construction theory of the gravity of (some) moral wrongs.” The idea here is not that, for instance, rape would be alright if it were considered alright; rather, the idea is that even from the perspective of a liberal morality—in fact: especially from the perspective of a liberal morality, given how important freedom and self-determination are for liberalism—there has to be leeway for a given society as to how it weighs different values and rights against each other, and what precise scope it gives to certain rights. For instance, liberalism does not see property rights as absolute,²² that is, one can take away part of a person’s property for tax purposes. But not only are property rights themselves non-absolute, there is also legitimate leeway when it comes to the social concretization of their scope: there is not the one absolutely right liberal taxing scheme.

Another example showing that the rights of people—even within a liberal framework—can be “socially constructed,” to use Leverick’s term, is the right to freedom. Liberals are very quick in saying that innocent people have such a right. Yet actual liberal societies endorse a practice that routinely deprives many innocent people of their freedom—namely the police practice of arresting suspects. The precise rules for this practice are rather complicated and differ from jurisdiction to jurisdiction. Roughly speaking, however, a police officer (without an arrest warrant) can make a lawful arrest of a suspect if she has “probable cause,” that is, if there is good reason to believe that the suspect has indeed committed a crime. Having good reason to think that someone has committed a crime is no guarantee that he did indeed. Often police officers arrest innocent people, even if they have probable cause. But if such an arrest of an innocent person violated the person’s rights, it would constitute an attack and thus trigger the applicability of a self-defense justification to resist the arrest. The suspect could say “no, I won’t comply,” and if the police officer tried to put handcuffs on him, he could push her back. If the police officer escalated her use of force further, the suspect could knock her down and even kill her as a last resort. Yet while quite a number of jurisdictions do indeed allow self-defense against unlawful arrest, including, in some cases, even lethal self-defense if necessary, no such courtesy is extended to the innocent victim of lawful arrest. (Sangero 2006: 132–136) To be sure, some authors [mistakenly (see Steinhoff 2016b)] think that “justification provides exemption from liability,” but they refer this to “objective justification.” (McMahan 2009: 43) Here, however, the police officer does not have an objective justification in the relevant sense. What justifies her is not that the arrest will objectively further the greater good or bring an actually guilty person behind bars; rather, what justifies her is merely the available evidence or her reasonable belief. Thus, the only way law and morality can plausibly deny the innocent suspect the right to resist lawful arrest is by denying him the right not to be arrested in the first place. This, of course, is precisely what law does, in the interest of safeguarding public peace and facilitating law enforcement. Morality, it seems, does exactly the same—unless one seriously thinks (as probably few will do) that in a reasonably just liberal-democratic state self-defense (and other-defense, for that matter) against a police officer conducting a lawful arrest is morally justified. Moreover, different jurisdictions also have different regulations regarding the speed with which a detention has to be validated or invalidated by a judge as well as regarding the length of permissible pretrial detentions. Claiming that among these variations there is the one right arrangement, while all the other variations deviate from the one clear path to liberal salvation, stresses credulity. Rather, it would appear that different liberal societies might balance the risk of placing an innocent person under arrest and the risk of letting a guilty person go free in different ways, and are, within limits, quite entitled to do so. However, if for the sake of upholding the social-legal order an even entirely innocent person’s moral right to freedom does not include a right against lawful arrest and often lengthy pretrial detention, why, then, should a guilty thief’s right to life necessarily include the moral right against being lawfully shot? It seems that if a society’s interest in effective law enforcement can outweigh even an innocent person’s interest in freedom and trim her rights accordingly, then society’s interest in the effective safe-guarding of property can outweigh a guilty person’s interest in not being killed while stealing from, and thus violating the rights of, others.

Thus, there is nothing illiberal—on the contrary—or objectionable in allowing a society to determine the precise scope of the proportionality requirement (if any) of self-defense itself, according to its own needs and its own balancing of different liberal rights and values against each other. And this will naturally affect the rights of the thief. Assume, for instance, that a society has adopted a rule according to which killing the thief in necessary defense of property is permissible. This rule imposes a significant risk on thieves of being killed while committing a theft. Would thieves have any legitimate complaint against the rule itself? It does not seem so. If a thief complained, “society” could answer²³: “Look, if you value the possession of a certain item so high and your life so low that you are willing to assume a significant risk of death to possess the item, you can hardly complain if we share your view: we also value possession of the item so high and a stealing thief’s life so low that we are willing to impose a significant lethal risk on the criminal activity of theft by endorsing a lenient proportionality requirement. So we entirely agree. What’s your problem?” That seems to be a good question indeed.

To be sure, the person killing the thief does more than merely impose a significant risk upon him. So does the thief have a valid complaint against him? No. In this case, if he has no valid complaint against the rule, that is, against the permission of lethal self-defense against a thief, he cannot have a valid complaint against the person acting on the permission, either. This is different in the case of a necessity or lesser evil justification. A necessity justification allows a person to infringe another person’s right for the greater good. For example, if an agent, at night, must break the window of a pharmacy to get the medicine who will save someone who has just collapsed in front of the window, he has a necessity justification for doing so. However, he would still have infringed the shop owner’s property right, as shown by the fact that he would owe the owner compensation for the broken window. What explains the difference is the fact that if you permit people to lethally defend themselves against a thief, you cannot simultaneously consider them as having violated the right to life—for this would then make them liable to pay damages often, or in most cases, exceeding the value of the property, thus defeating the purpose of the rule. Such an arrangement would not permit lethal self-defense against theft but in fact penalize it by making it exceedingly costly. Conversely, if the compensation due for the alleged violation of the right of life of the thief is considered smaller than the value of the property defended, on what grounds could one then rule out the permissibility of defending one’s property against him in the first place? Even according to Leverick’s forfeiture theory, the thief would then actually have forfeited his right to life by posing a threat to something of greater value.

Thus, there seems to be no way around the conclusion that the permission to engage in lethal self-defense can only go hand in hand with the thief’s not having a right against such lethal defense. Whether or not he does have a right, however, depends in large part on whether his society recognizes such a right. By saying this I am not endorsing a full-fledged recognition theory of rights, according to which a necessary condition for the existence a right is that it is socially recognized²⁴; rather, I am endorsing, at least in the context of the proportionality condition, a recognition theory of the scope of certain rights, including the right to life, and I have tried to demonstrate that this recognition theory also makes a lot of sense with regard to other rights, like the right to liberty (my example were arrest rules and rules of pretrial detention) and the right to property (example: taxation). Even a deontologist relying as much on what can be determined “a priori” as Kant, after all, acknowledges that the precise scope of property rights have a social element. (Kant 1996: 44–46) And the legal scholar Paul Robinson (1984: 87) notes, with regard to the alleged retreat rule, which he rightly discusses under the broader heading of the proportionality requirement: “Like other judgments that are appropriately left to legislative decisions, the ultimate conclusion will vary with the differences in different communities’ balance of interests.” However, the “legislative decisions” themselves will often simply reflect the society’s understanding of the scope and limits of the right, and this understanding is morally relevant. One cannot know what the morally adequate proportionality requirement of a society is if one simply ignores the society’s balance of interests and its understanding of the right.

Thus, the fact that some liberal societies have meanwhile opted for proportionality requirements that are comparatively lenient on aggressors is hardly able to show that less aggressor-friendly proportionality requirements are immoral, illiberal, or rights-violating. There is simply no reason to assume that it is necessarily disproportionate to kill in defense of property or minor assaults. Instead of reasons, all defenders of the contrary position usually do is insist on their “bedrock” intuitions that it is disproportionate and on rhetorical appeals to what they consider to be “civilized society.” Ironically, however, some of these defenders are then undermined by their very own theories: Leverick, for instance, by her social construction theory of the gravity of (some) moral wrongs, and Rodin by his reciprocity theory of rights. Instead of bedrock, we get sand. In contrast, I did provide reasons why lethal defense of property does not necessarily violate the thief’s right to life, and these reasons cohere, as far as I can see, with the rest of my theory of self-defense.

Proportionality as Precautionary Rule

In the previous section, I argued that there is not necessarily a right not to be killed in defense against theft or minor assaults. At the beginning of that section I also said, however, that proportionality restraints need not be based on the rights of aggressors alone. This parallels the case of the necessity requirement; yet the case of proportionality is more complicated. It is perhaps best to approach the proportionality case by first considering the case of necessity.

Suppose Frank unjustly tries to kill you simply because he does not like your nose. You can as easily stop him by using your taser as by shooting him in the chest with your revolver. You shoot him in the chest. On my view, this is unjustified (and in a moment I will explain why.) However, have you wronged him, that is, can the aggressor justly complain? Did you violate his rights? I think one can reasonably deny that. After all, if the aggressor says (dying), “You used lethal self-defense although you could have used less drastic means,” the understandable and quite appropriate reply would be: “Look who’s talking.” (See also Frowe 2013: 112–113) It seems that by culpably and unnecessarily trying to kill you the attacker has forfeited his right against you that you not kill him unnecessarily.²⁵

This is different in the case of an innocent aggressor, for example in the case of someone whose drink has been spiked with a hallucinogenic drug, so that he now takes you for the hostile extraterrestrial from Alien.²⁶ Innocent aggressors can stand to complain if they are unnecessarily harmed, and thus they do retain their rights against unnecessary harm (at least if they would have shown the same restraint if you had innocently attacked them). “Look who’s talking” is not an equally apt reply to both Frank and an innocent aggressor. Whether people are responsible actors or not obviously matters a great deal for moral questions. This is also shown in the fact that the innocent aggressor could justly respond: “What do you mean, ‘Look who’s talking’? You reproach me with not doing to others as I would have them do to me? But you are wrong. If I had been of sound mind and you had attacked me under the influence of drugs, I would not have used unnecessary force to stop you.” Provided this is the honest truth—and of course it can be—this is a valid rejoinder. This rejoinder, however, is not available to Frank. He was of sound mind (in the sense that he was a responsible actor) and nevertheless did use unnecessary force even without anybody attacking him. Thus, if he were to claim that he wouldn’t have used unnecessary force against you if you had tried to kill him, then this would be an obvious lie.

Abiding by the rule not to harm aggressors unnecessarily will reduce the likelihood that the defender unnecessarily harms innocent people and thereby violates their rights. Insofar as morality requires agents not to impose unfair and unreasonable risks on others, it requires defenders to abide by the necessity requirement. This requirement is thus indirectly based on rights (though not of aggressors but of innocent people), but violating it is not the same as violating a right. The rationale is also not an overall “rights-utilitarianism” that tries to reduce overall rights-violations: rather the rationale is an agent-relative requirement to take reasonable and fair precautions to reduce one’s own likelihood of violating other people’s rights. Thus, the necessity requirement is (at least also) a precautionary rule.

So far the argument for the external (that is, external to the rights of aggressors) necessity requirement. This requirement is, in my view, almost absolute, because there is always a risk, however small, that one is confronted with an innocent aggressor. The requirement can sometimes be overridden, but then we leave the ambit of the self-defense justification and enter the ambit of the lesser evil or necessity justification and of punishment.²⁷

As already noted, the situation with proportionality is more complicated. It is more complicated precisely because proportionality considerations involve the weighing and balancing of different interests. Here, there are no absolutes. Consider, for instance, a society that has been faced with rampant theft and in “societal defense,” to put it this way, developed proportionality rules of self-defense that do, in principle, allow killing thieves. However, it is also realized that sometimes innocent people get killed: people who have been forced by other criminals to steal and thus acted under duress, insane persons, and children. And sometimes people get killed in putative defense against theft because the “defender” made a mistake and misinterpreted the situation.

Thus, society is forced here to weigh different interests and risks (for example of making mistakes) against each other. The weighing moreover is not just of property against life or bodily injury. It involves rights, a person’s sovereignty over her domain, honor (a victim of theft is of course also disrespected, and she might consider it dishonorable not to defend herself), and the social-legal order.

In that context it should also be noted that the social-legal order is undermined not just by culpable aggressors but also by innocent ones. Boaz Sangero sees this differently, claiming that the opposing German view displays a “strange perception.” He opines that “in the absence of criminal responsibility, the aggressor’s action bears no anti-social nature, and accordingly it does not—certainly not significantly—injure the social-legal order.” (Sangero 2006: 71, n. 345) I think that it is actually this perception that is strange. I submit that the German understanding of the social-legal order is much less abstract—and much more plausible—than Sangero’s. Imagine, for instance, that you are in a mess hall, and everyone behaves as they should, standing in line, moving forward when required, taking their seat in an orderly fashion. Suddenly five officers jump up, draw their sidearms, and start shooting at everybody. Tumult breaks out, people take cover, blood is spattered around. It seems that if you were in this situation, you would think that the social-legal order in the mess hall has broken down, and you would hardly change your opinion if you later found out that the five officers had a psychotic break due to drugs that had accidentally been put into their food. “Oh, they were innocent? In that case the order in the mess hall has never been seriously undermined by their shooting and killing us, I was wrong.” No, you weren’t. Likewise, in a society where people constantly carry away other people’s property, the social-legal order is seriously undermined, and hardly much less seriously so if it turns out that those carrying away the property are insane.

As I said, the weighing to be done is difficult, and each society has to find the proper balance for itself. To look from our society at a society like, for instance, the Old West, where horse thieves were punished by hanging or their thievery thwarted by use of lethal force, and to then judge that these practices are disproportionate and “barbaric,” might get a warm-hearted appreciative nod in certain circles in our society, but it also seems rather context-insensitive, unrealistic, and naïve. “What’s right in our society is right for others”—not really. It should be noted that in a society where horses have huge importance, where official law-enforcement is ineffective, and where therefore the use of lethal force against horse thieves is legal, it is highly likely that the horse thieves themselves would defend their horses by the same measures. If so, they are hardly in a position to complain if people do to them what they would do to others. Moreover, given the balance of—liberal—interests the society finds appropriate, the horse thieves, for reasons expounded above, do not even have standing to complain if they themselves would not use lethal violence. After all, analogously, if someone in my society tries to make off with my coffee, I am also not violating his right if I keep him from doing so by grabbing his arm—that he wouldn’t have grabbed mine if the situation had been reversed does not change this.

Conversely, however, the fact that a society recognizes the right of thieves not to be killed in mere defense of property can create such a right. (Likewise, if a society recognizes, for example through its legislation, a right of 16 year old persons to vote, then this also creates a moral right of those persons to vote, a right that they might not have had before.) If the thief is killed under such circumstances, he now does have standing to complain—unless, again, he would, had the situation been reversed, have killed the thief himself. In that case, killing him in defense of property would not violate his right. Yet it would still violate society’s proportionality rule, which expresses society’s balancing of interests. The defender, by violating this rule, has taken a disproportionate risk of killing someone who did have the right not to be killed. Therefore the killing, while not a rights-violation, would still be unjustified.

Conclusion

In this paper, I first laid bare the assumptions and the logic—and often illogic—underlying very strict accounts of the proportionality requirement. I argued that accounts that try to rule out lethal self-defense against threats to property or against threats of minor assault by an appeal to the supreme value of life have counter-intuitive implications and are untenable. Furthermore, I provided arguments demonstrating that there is not necessarily a right not to be killed in defense against theft or minor assaults. While there is a general moral right of self-defense and a general right to life, the scope of these rights (like the scope of the right to liberty and the scope of the right to property) depends on certain social facts that—even within a liberal framework—can differ from one society to another. Moreover, the proportionality of self-defense does not depend on the rights of the aggressor alone, but also on a precautionary rule, shaped by the balance of interests of the society in question and aimed at protecting innocent people and other social interests. This rule can protect an aggressor even in cases where he does not have the right to such protection.