Jump to ContentJump to Main Navigation
Targeted KillingsLaw and Morality in an Asymmetrical World$

Claire Finkelstein, Jens David Ohlin, and Andrew Altman

Print publication date: 2012

Print ISBN-13: 9780199646470

Published to Oxford Scholarship Online: September 2012

DOI: 10.1093/acprof:oso/9780199646470.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 12 December 2018



Targeted Killings

Claire Finkelstein

Oxford University Press

Abstract and Keywords

The policy of the Bush Administration with regard to terror suspects focused on detention and interrogation: detention as a temporary immobilization of potential terrorists, and interrogation to gather information needed to fight the war on terror. No one was surprised, then, when the new Obama Administration wanted to distance itself from the tactics of the former Bush Administration. It sought ways of addressing the threat of terrorism that avoided the brutality in interrogations, indefinite detentions, and the moral and legal thicket of conducting legal proceedings against terror suspects, whose situations were dramatically different from ordinary criminal defendants. It is perhaps no accident, then, that targeted killing emerged as the central strategy for fighting the war on terror. This chapter begins by arguing that the justification that applies to the practice of killing in war, in its traditional form, cannot properly be extended to the practice of targeting previously identified individuals in a way that abstracts from the proximity of their connection to active hostilities. The practice of targeted killing, as currently fashioned, is for this reason not easily justifiable under the traditional laws of war. It suggests however, that there are yet other rationales for killing those who pose a danger to national security that do not rely on either traditional just war theory or on the domestic law of personal justifications. The chapter advances an argument for an expansive approach to prevention called ‘preemptive killing’. Preemption, unlike prevention, extends the preventive privilege to a number of cases in which the anticipated harm is non-imminent.

Keywords:   war on terror, President Obama, President Bush, targeted killing, justification, preemptive killing

I. Introduction

The policy of the Bush Administration with regard to terror suspects focused on detention and interrogation: detention as a temporary immobilization of potential terrorists, and interrogation to gather information needed to fight the war on terror. With emerging public awareness of the inhumane conditions of detention and the harshness of our so-called “enhanced interrogation” techniques, our practices in the war on terror had become a stain on America's conscience and a source of international embarrassment.1

No one was surprised, then, that the Obama Administration wanted to distance itself from the tactics of the former Administration. It sought ways of addressing the threat of terrorism that avoided the brutality in interrogations, (p.157) indefinite detentions, and the moral and legal thicket of conducting legal proceedings against terror suspects, whose situations were dramatically different from ordinary criminal defendants. It is perhaps no accident, then, that targeted killing emerged as the central strategy for fighting the war on terror. The new policy involved a dramatic shift in operations.2 To look only at the statistics for targeted killing operations undertaken by drone, for example, between the years 2004 to 2008, the Bush Administration authorized 42 targeted killings, by comparison with the Obama Administration's count of 180 authorized drone strikes as of February 11, 2011.3 BetweenFebruary and the time of this writing (November, 2011), the number of drone strikes has further increased significantly.4–5

While there has been a steady increase in reliance on targeted killing as a technique of war, the willingness on the part of the Obama Administration to subject the new policy to the legal and moral examination it appears to warrant has not kept pace. There are signs that the Administration has embarked on a campaign of targeted killing with the same unreflective enthusiasm that the Bush Administration displayed with regard to the use of torture to aid interrogations. To be sure, it is not difficult to understand the attractions of the practice: it allows the more unsavory aspects of the war on terror to be sanitized and removed from public view. But perhaps more legitimately, it allows the United States to avoid the problems associated with amassing large numbers of detainees whose captivity is difficult to manage or justify under principles of international law, and who cannot easily be brought into any court of law for trial. As law professor Ken Anderson has suggested: “Since the U.S. political and legal situation has made aggressive interrogation a questionable activity anyway, there is less reason to seek to capture rather than kill . . . . And if one intends to kill, the incentive is to do so from a standoff position because it removes potentially messy questions (p.158) of surrender.”6 Anderson's point seems hard to deny: if our interrogation and detention policies were typically conducted at the edge of, or beyond the bounds, of what is legally and ethically permissible, does it worsen our moral position to kill without attempting to capture? Might not targeted killing be thought the lesser of two evils?

Yet the turn away from interrogation and detention towards targeted killing has not succeeded in removing the war on terror from political controversy in the way the Obama Administration may have hoped.7 On the contrary, it now appears that killing terror suspects in lieu of detaining them does not eliminate the myriad difficulties the world witnessed with capture and detention. The problem has just been moved to a different spot under the rug. The operations involved in combating terrorist activity do not fit squarely within the principles of justification required by just war theory, and this has created legal and moral dilemmas in many domains simultaneously. We have sometimes attempted to solve those difficulties by relabeling the key distinctions on which just war theory depends, but this has not resolved the deeper legal challenges posed by the war on terror.8

The most serious of the conceptual difficulties the new style of warfare faces is the fact that in a war waged against terrorists and civilian militants, it is not clear who, if anyone, should count as a “combatant.” This fundamental indeterminacy renders obscure who can be legitimately targeted, who can be detained, as well as the justification for targeting or detaining, the extent of the duty to seek capture before killing, and, more generally, whether detainees in the war on terror should fall under the protections traditionally extended to prisoners of war. The lack of theoretical clarity in this area allowed Bush Administration lawyers to interpret any applicable legal constraints “loosely,” namely in a way that was designed to enhance (p.159) the legitimacy of the Administration's interrogation practices without excessive focus on the traditions and laws of war, as well as with international humanitarian law.9 In particular, an enduring legacy of these efforts was the adoption of a third legal category, one intended to identify the central figures in the war on terror as targetable, like combatants, but exempt from the protections afforded co-belligerents under Article IV of the fourth Geneva Convention and under traditional just war theory.10 These were the so-called “unlawful” combatants. In this way, the Administration was able to avoid the constraints of the Geneva Conventions’ protections for prisoners of war, but also the Federal Torture Statute,11 the Convention on Torture (CAT),12 and other national and international rules governing the trial and detention procedures for captured enemy combatants in dealing with members of Al Qaeda. The idea of treating members of Al Qaeda as non-enemy combatants was of course the central legal rationale for the former Administration's interrogation policies. Without this convenient middle tier, our treatment of terror suspects would likely have proven yet more difficult to defend.

These same aims have induced the Obama Administration to retain the Bush legal structure with regard to suspected terrorists. The in-between category has been as significant for justifying the policy on targeted killing as it was for the proffered justification of enhanced interrogation techniques. The legal and moral obstacles to legitimizing enhanced interrogation of non-enemy combatants now arises with respect to the legitimacy of targeting individuals who are arguably civilians, but whose contribution to the terrorist objectives of Al Qaeda have increased the reach and danger of the primary actors in the war on terror. The recent killing of Anwar al-Awlaki provides a clear example. If al-Awlaki's activities in support of Al Qaeda were propagandistic rather than actively belligerent,13 he probably should not have been regarded as a “combatant” under principles of traditional just war theory, and for purposes of the Geneva Conventions, and hence the legitimacy of killing rather than capturing him would be subject to doubt. Similar concerns arise with regard to whether the constraints that operate on military interrogators also apply to members of the executive branch, such as the CIA, and whether it is legitimate (p.160) for non-uniformed executive officials to engage in targeted killing.14 The dilemmas are the same; only their expression has changed.

In both of these morally and legally fraught areas of policy we essentially find ourselves caught between two paradigms of justification: the basic paradigm of killing in war, which depends crucially on the distinction between combatants and civilians, on the one hand, and a more generic framework for justifying the use of violence, such as grounds the entitlement to use force in a law-enforcement context. The first paradigm is a specialized one: the available justifications for killing in war are domain-specific, meaning that they apply to war and to no other domain. The distinction between the jus ad bellum and the jus in bello captures this idea. While the former domain is open to all of morality, the latter brackets general moral concerns, in favor of arguments and rationales that apply narrowly to the context of aggression between co-belligerents. The second paradigm is open to general arguments of morality and is not domain-limited. Its logic pervades general moral practice, as well as the basic approach to the concepts of justification and excuse in the criminal law. The latter domain is highly reflective of moral practice, and thus studying the structure of the criminal law will provide us with something like a template for the demands of ordinary morality regarding judgments of responsibility.15 While it is not clear that either will ultimately succeed in rationalizing the use of enhanced interrogation techniques or the targeting of noncombatants, these two sources of possible justification exhaust the moral resources we have at our disposal to address the ethics of these marginal practices in war.

The practice of targeted killing is conducted by military personnel or executive branch officials according to a policy of tracking down and killing individuals whose names appear on a classified kill or capture list known as the Joint Prioritized Effects List (JPEL). An individual's appearance on the list is based on his importance to the noncombatant enemy force, as well as on the degree to which he threatens domestic interests. Suitability for the list, however, is not exclusively a product of the imminent danger the “target” is taken to pose, as would be the case were the justification for killing limited to self-defense, defense of others, or law-enforcement. According to a former U.S. Foreign Service officer, “the list included bomb makers, commanders, financiers, people who coordinate the weapons transport (p.161) and even [public relations] people.” The list has been described as part of “an almost industrial-scale counterterrorism killing machine.”16

What is the justification for the practice of targeted killing, as described above? Let us ask the question of justification from the standpoint of what Michael Walzer calls “the war convention,” namely the combined set of moral principles and background set of legitimating norms that make up what has traditionally been called “just war theory.”17 Justifications for the practice must belong to one of two categories: either the practice is justified killing of co-belligerents, as set out by the traditions and laws of war, or it is a form of law-enforcement, whose norms are established by the parameters of the general principles of morality relating to justifications and excuses. If the practice of targeted killing cannot be justified under one or the other of these categories, it cannot be justified at all. As Michael Gross writes:

Either soldiers are criminals or they are not. If they are not outlaws, then there is no cause to declare them criminals or kill them covertly. If they are criminals, however, then they should be charged, arrested, tried and sentenced, not shot on sight. Killing criminals without the benefit of trial smacks of extrajudicial execution.18

Our current approach to targeted killing is betwixt and between. We treat targeted individuals as belligerents insofar as we regard them as legitimate targets by virtue of status, rather than action. But we treat them as subjects of law enforcement in that we resist according them the privileges that go along with the status of combatants, such as affording them the rights of P.O.W.s and recognizing their equal right to kill in combat.

As is the case with all intentional killing, in the absence of an affirmative justification, targeted killing is morally impermissible. If an actor kills intentionally and lacks a reason that could justify his actions, he must be called to account. He may turn out to be exempt from responsibility by a personal defense, such as insanity, infancy, involuntary intoxication, duress, or for other status-based reasons. He may even in some cases be able to name the fact that he was following orders as an excuse, rather than a justification, in cases in which he made a reasonable mistake about the legitimacy of the orders.19 This chapter will consider the prospects for (p.162) justifying targeted killing according to the foregoing paradigms, and will ultimately suggest a justification for the practice that falls, roughly speaking, within the law-enforcement model, according to a somewhat attenuated version of that basis for legitimizing the use of force.

I shall begin by arguing that the justification that applies to the practice of killing in war, in its traditional form, cannot properly be extended to the practice of targeting previously identified individuals in a way that abstracts from the proximity of their connection to active hostilities. The practice of targeted killing, as currently fashioned, is for this reason not easily justifiable under the traditional laws of war. I shall suggest however, that there are yet other rationales for killing those who pose a danger to national security that do not rely on either traditional just war theory or on the domestic law of personal justifications. One such alternative can be found in an extension of the privilege to prevent the commission of rights violations against one's person or the person of another whom one is entitled to defend. Although as applied to the practice of targeted killing, the justification will not turn out to be a traditional preventive rationale, I shall advance an argument for an expansive approach to prevention I call “preemptive killing.” Preemption, unlike prevention, extends the preventive privilege to a number of cases in which the anticipated harm is non-imminent. Preemptive killing, however, is also more limited than either preventive killing or killing in accordance with just war theory, and thus its scope needs to be carefully identified and its application sharply circumscribed in accordance with its background justification.20 The practice that emerges as justified on this account is somewhat different from the use of targeted killing as currently practiced. Normative theory as applied to national security practices in the war on terror thus suggest a re-examination of current policy. I shall then address the pragmatic implications of the theory of justification I present.

II. Targeted killing and the realities of modern warfare

Even more than the legal and moral soul-searching raised by the practice of detaining so-called “non-enemy” or “unlawful” combatants, the practice of targeted killing, and its perceived role in judgments of military necessity, casts in relief the complicated realities of modern warfare. This is in significant part a reflection of the degree to which the practice of targeted killing departs from the traditional battlefield form of combat, and hence from the core justifications for killing in war. There are four distinct ethical concerns that targeted killing, as currently (p.163) practiced, raises. I shall identify them briefly here, and then proceed to a more detailed discussion of these concerns in what follows.

First is the fact that individuals “targeted” according to a policy of targeted killing are named in advance of the attempt on their lives. The authorization to kill or capture them does not hinge, once they are on the list, on any particular behavior on their part. The relevant list is the JPEL and this is amended weekly and drawn up with intelligence input from the National Directorate of Security and the U.S. Central Intelligence Agency.21 The authors of the list then assign serial numbers and code names to the identified suspects, who become the objects of intensive research in order to determine their patterns of behavior and to maximize the chances of killing (or in theory capturing) them in a planned operation. Those on the JPEL eventually become the targets of a Predator drone attack, a strike by F-15E strike aircraft or, in the last resort, a night-time assault by a TF 373 mission dispatched from a local military base.22

That the killing is “targeted” is not the issue. On the contrary, that killing is directed towards a target rather than randomly fired would appear to be a sine qua non for justifying it. The alternative, namely random killing, is hardly a practice worth seeking to defend. The problem is that the list involves the advance naming of individuals for targeting, a practice that presents justificatory problems. Commentators have mostly overlooked this aspect of the practice, and their criticisms have focused not on the individual identification of those targeted, but on the mere fact that individuals are targeted at all. It is important to notice, however, that there is a significant moral difference between targeting anyone who satisfies certain generic criteria, and targeting a particular person on the basis of the fact that he meets those criteria.

Second, a quite independent concern from advance naming is the relative expansiveness of the criteria for inclusion on the list of named suspects. Unlike where killing in war is concerned, the practice of naming and targeting terror suspects extends beyond those who are engaged in active combat.23 The question then arises whether it is legitimate to target individuals involved in the War on Terror who lack the status of active combatants, and what the outer reaches of the criteria for inclusion on the list of those who can be targeted should be. Without a clear theory of the justification that lies behind targeted killing operations, we cannot begin to identify the legitimate scope of the practice. This is a significant lacuna in the (p.164) jurisprudence of the war on terror, because the relationship between the standard act of killing in war and the targeted killing of quasi-combatants, “non-enemy combatants,” or engaged and active bystanders has not as of yet been adequately theorized.

Third, if we allow that targeting at least certain individuals who are instrumental to the war on terror is legitimate, the further question arises of who may engage in the killing. If targeting those with a more attenuated relationship to active combat does not strictly speaking constitute an act of war, does that have implications for who may legitimately engage in targeting? Does relaxing the rules on the identity of the target imply a similar relaxation of the rules on the identity of the killer? The difficulty is that allowing CIA officials to conduct targeted killing, while a natural extension of many of its past practices, seems tantamount to hiring civilians to engage in acts of war, albeit executive branch officials. As Gary Solis has argued, allowing CIA officials to operate unmanned aerial vehicles makes them “civilians directly engaged in hostilities,” an act for which they could themselves be branded “unlawful combatants” and subject to prosecution.24 He writes:

CIA civilian personnel who repeatedly and directly participate in hostilities may have what recent guidance from the International Committee of the Red Cross terms “a continuous combat function.” That status, the ICRC guidance says, makes them legitimate targets whenever and wherever they may be found . . . While the guidance speaks in terms of non-state actors, there is no reason why the same is not true of civilian agents of state actors such as the United States.25

As for the arguments in defense of the practice that CIA officials can constitute belligerents assisting the military, the practice arguably flies in the face of a time-honored constraint on the right to kill in war: that belligerents identify themselves with visible insignia.26 When combined with the second point, namely that it is not always active combatants that are placed on the “kill list,” we potentially have an official governmental policy of hiring civilians to target other civilians—a far cry from the traditional reciprocity conditions and the principle of distinction that have been central to the maintenance of ethical standards in war. Thus although (p.165) the practice of targeted killing is derived from a model based on combatant killing in war, the practice is a significantly attenuated version of that model, and hence merits careful legal and moral scrutiny. Ultimately what would be needed to justify the practice in its current form is a reconceptualization of the relevant portion of just war theory to take account of the “asymmetrical” nature of modern warfare. Whether such a reconceptualization is possible within the constraints of the justificatory constraints that typically apply to intentional killing is a serious question, one whose answer should not be taken for granted.

Fourth, the practice of targeted killing has become associated with the use of drones in order to effectuate the assassination of individuals on the named list. As Ken Anderson rightly points out, this is at least in part because drone killing minimizes the risk that the target will attempt to surrender and convert himself into a detainee.27 The question that arises, however, is whether the justification for targeted killing, once identified, should be understood as placing any restrictions on the method or form such killing takes. While commentators often express ambivalence about the use of drone technology,28 whether it is ultimately justified depends on several crucial factors. Does the “target” have a right under the laws of war to surrender if affording him this opportunity is militarily feasible?29 If he has the right, does the use of drones make it significantly more difficult for him to exercise that right? From the standpoint of the duty to minimize civilian casualties, there may be an obligation to treat drone killing as a last resort, given that it unduly escalates killing, and thus draws international conflict further and further away from a model of lawful, symmetrical killing in war.30 Because these fundamental questions in just war theory have not been answered, we do not know how to assess the moral status of using drone technology in targeted killing. The reverse is also true: the failure to resolve issues connected with problems at the margins of just war theory, such as the use of drone technology, has a reverse inferential effect on the more standard cases of killing in war. If the justification for using drones or for killing without affording an opportunity to surrender is not clearly established, the practice calls the most fundamental premises of just war theory into question. The moral and legal ambiguities of modern “targeted killing” are thus a reflection of the profound ambiguities of modern warfare, which press at the boundaries of the traditional ways in which war has been conceived, and with the moral justification for intentional killing that has historically been offered as accompanying the reciprocal approach to war. For this reason, the traditional paradigm, represented in just war theory as a conflict between enemy soldiers (p.166) attacking and counter-attacking in a clearly defined physical space of combat, has only a tenuous application to the moral problem of justifying lethal acts in the War on Terror.

In what follows, I shall restrict my focus to the first and the final aspects of the current practice of targeted killing. While I will make reference to the other two problems, it is the first and the fourth problems that most clearly distinguish targeted killing as it is currently practiced from other practices in war, and they are also the aspects of that practice that most cast distinctive doubt on its legitimacy. If we are able to justify the practice of targeted killing in the face of these two significant challenges to its legitimacy, we will have addressed the central objections to the use of targeted killing, and in this way substantially answered the concerns recent leveled against the practice.

III. The legitimacy of naming targets

Let us begin by focusing on the concerns I raised in the previous section about the practice of naming subjects to be targeted. What difference does it make whether subjects are identified by name on a “kill or capture list”, or whether the decision to identify someone as a target is based on actual threatening behavior? There are two reasons to object to the advance naming practice from an ethical standpoint, and these differences with the standard form of killing in war suggest significantly increased difficulties finding a justification for the practice in its current form. The two problems are, as I shall call them, first, the bootstrapping problem, and second, the problem of statistical versus identified targets. First, the bootstrapping problem.

The question has recently arisen whether it would have been permissible for the United States to target Colonel Muammar Qaddafi, who as it happens met his demise at the hands of Libyan National Transition Council (NTC) forces after fleeing his convoy which had been attacked by NATO warplanes and a U.S. drone.31 On the one hand, he was, by his own admission, responsible for terrorist attacks against the United States, in particular the bombing of Pan Am Flight 103 over Lockerbie, Scotland in 1988.32 He continued to support terrorism, and he was a sworn enemy of the United States and of his own civilian population. Surely the world is a better place without him, almost no matter what (p.167) arises in Libya to fill the void he has left. On the other hand, it is not clear under the laws of war whether the foregoing considerations are relevant. Whatever his crimes, Qaddafi was a sitting head of state, recognized by various groups as a legitimate ruler, and as such possessed certain rights of sovereignty that cannot be ignored with respect to any foreign leader.33 There is little justification under the laws of war for assassinating a head of state on the ground that he is guilty of a crime against the United States, for which he could potentially be tried in the International Criminal Court, or because we see him as a tyrant to his own people and may be guilty of crimes against humanity on multiple occasions. The only legitimate basis for targeting him in the absence of an existing state of war between the United States and Libya would have been that we had reason to fear his instigation of imminent attacks against the United States through terrorism or otherwise that could not be prevented by lesser means.

Matters would have been different, of course, if we had been war with Libya, as was the case with Saddam Hussein. Qaddafi was the commander of the armed forces and therefore by definition a part of Libya's combat forces in times of war.34 That enemy forces would be entitled to kill him under these circumstances is a non-controversial byproduct of the theory of war, at least to the extent that killing him can be deemed militarily necessary. It is, of course, a further question whether they are entitled to give up an opportunity to capture him in order to kill him instead, if they are in a position of being able to do either (and either would serve the purposes of military necessity). One would have thought the moral principle outside the context of war that dictates using the least amount of force necessary to satisfy the ends of survival would govern this case, but that added dimension of the debate about targeted killing is one I will leave to one side for the moment.35

There is, however, a rather more controversial aspect of the hypothetical targeted killing of Qaddafi we are considering. Had we attacked Qaddafi we would have been at war with Libya, and that might appear to justify treating him as a combatant and targeting him, given his position as head of the armed forces. But presumably in that case, we ought not to regard a war with Libya as a neutral fact, something that just happened all by itself. We are the ones in this scenario, after all, who would have initiated acts of war against Libya. To treat such acts as justifying targeting Qaddafi as a combatant seems like a piece of morally specious bootstrapping. It is, indeed, reminiscent of the doctrine of the Actio Libera in (p.168) Causa, otherwise known as creating the conditions of one's own defense.36 Consider the following case.

If I want to kill my enemy, but do not want to be guilty of murder, one thing I could do is to induce him to attack me with the advance plan of killing him in supposed self-defense. Most criminal codes will, however, deny a defendant a self-defense claim if “the actor, with the purpose of causing death or serious bodily injury, provoked the use of force against himself in the same encounter.”37 Similarly, the defense of duress is generally unavailable “if the actor recklessly placed himself in a situation in which it was probable that he would be subjected to duress.”38

Intuition speaks strongly against allowing the ground schemer the defense in such situations, though the point remains controversial. Would Qaddafi's status as a “belligerent” have been a product of our own attacks o Libya, and if so, would this be a basis for denying the traditional justification for killing co-belligerents in war?39 The concern about the application of the reasoning of the Actio Libera in Causa in this context makes clear that there are at least some situations in which the justification for targeting a given individual may depend in a crucial way on why he bears the status of belligerent with respect to the United States, thus effacing to some degree the line between the jus ad bellum and the jus in bello.40 At the very least, a total severance of ad bellum reasoning and in bello reasoning would allow for the following sort of abuse.

Imagine we would like to target a head of state of a humane democracy. We know placing such an individual, who is a nonbelligerent, on the list of targets is impermissible under the laws of war. According to the above logic, we have only to declare war on that country in order to convert the relationship into one between belligerents, and although such a declaration would be impermissible from an ad bellum standpoint, it would entitle us to target the now-enemy head of state without subjecting ourselves to liability for war crimes. As long as the in bello criterion for the legitimacy of killing enemy combatants does not depend on the justice of (p.169) our cause in declaring war in the first place, we can kill with impunity as long as we have committed the quite different wrongful act of engaging in a wrongful declaration or act of war. The final step in this hypothetical is to notice that wars are rarely “declared” these days, and we do not generally consider the onset of war to be a violation of ad bellum constraints just because it was not declared. If this is correct, then it seems we need only perform an act of combat in order to turn ourselves, and our opponents, into combatants. And this requirement is satisfied by an attempt to assassinate the sitting head of state of the other country! So it seems we can bootstrap our way up into establishing the legitimacy of our own wrongful acts of targeting by engaging in illegal-targeting-turned-legal by its own existence. The constraints formerly posed by the law of war here do no work towards establishing normative constraints on the act of killing in war.

The bootstrapping scenario is of concern in its own right. But it is of particular relevance here, because it makes clear the potential moral risks in status rather than conduct, based killing. Naming a target in advance, and then killing him based on his status, requires that he fall into the category of combatant if such actions are to be permissible under principles of just war theory. But many targets we currently name, perhaps most of them, cannot be justifiably identified as status-based targets without relying on a distortion of the category of “combatant.” Such distortions, in effect, stem from a failure to adequately distinguish jus ad bellum from jus in bello norms. Lack of clarity about what constitutes a legitimate ground for war then translates into a weakening of the traditional conduct requirements for justly prosecuting a war. The bootstrapping phenomenon thus arises out of the same decision to ignore the age-old principle of distinction between combatants and civilians. The principle of distinction must be significantly attenuated before we can seek to justify killing members of a non-governmental terrorist organization as combatants: unlawful combatants in the War on Terror are treated like combatants for the purpose of claiming the legitimacy of using status-based justified killing against them, but we treat them like civilians with regard to their own combatant and detainee privileges. The second questionable aspect about naming nonbelligerents (or self-created belligerents) in advance of an actual attack based on their current conduct has to do with the difference between statistical and previously named victims, a moral phenomenon of much more general applicability that nevertheless appears to play a role in this context. Consider the following example.

Suppose a real estate developer is trying to decide whether to proceed with the building of a large skyscraper in the middle of a busy downtown area. Imagine two possible scenarios. Scenario One: the developer knows to a very high degree of likelihood that at least one person will die in the construction of this building.41 (p.170) He will of course take all precautions against the materialization of this risk, but he must consider the “cost” of going ahead with the project as containing some forecast of either wrongful death compensation or of payment for injury and/or disability. In general, we accept the permissibility of welfare-enhancing projects that bear negative externalities, as long as those projects remain socially productive once externalities are internalized. A foreseeable risk of social loss does not leave a moral stain on high-risk construction, any more than foreseeable risk of a certain level of road fatalities leaves a moral stain on driving.

Consider, however, Scenario Two. The developer this time is aware not only that there is a high degree of certainty that someone will die or become severely injured in the building of the sky scraper, but that Fred, one of his workers, will be the one to succumb to this loss. Is it permissible for the developer to proceed with the construction project in the face of the known or anticipated loss of an identifiable member of his work team? Philosophers and legal scholars generally treat risks that fall in the first category, which they call “statistical risks”, as acceptable to run, but regard risks to identifiable individuals, which we might call “named risks”, as wholly different in character. Why is not clear, but the intuition tends to be strongly felt. It is particularly odd, given that statistical risks can be run over a group of individuals all of whom are known to the risk taker, and yet the sense is that statistical risks are acceptable under certain conditions, but that named risks never are.

It seems reasonable to suppose that the discomfort one might feel with the fact that targeted killing involves the prior identification of named targets may be comparable to the different reactions we tend to have to statistical versus named risk. Now it is true that the imaginary example of building a skyscraper knowing exactly who is likely to be killed when we undertake the work is different in important ways from the scenario of targeted killing we are attempting to assess. Most importantly, the skyscraper scenario involves exposing human beings to a foreseen, rather than an intentional risk of death. Death, in both the statistical and the non-statistical versions of the skyscraper example, is a side-effect of an otherwise productive activity. But notice which way this argument cuts. If one cleaves to the moral significance of intention, then targeted killing should be more difficult to justify than building skyscrapers, even knowing that a particular person, Joe, is highly likely to be killed in the procedure. Since intentional harm is thought by many to be more difficult to justify than incidental or foreseeable harm, then a fortiori if we would not proceed in the skyscraper case having identified the victim, we ought not to proceed in the targeted killing case, where the victim is comparably identified. The intuition that it is impermissible to build the skyscraper in the named case, then, does have implications for our targeted killing practice insofar as the latter involves advance naming, rather than mere statistical identification. (p.171)

To better understand the intuition at work here, consider the constitutional prohibition on bills of attainder, in which individuals were previously identified as meriting judicial exile, and their status as legal persons was systematically eliminated, both through criminal punishment and through escheat of their possessions to the state.42 While I do not have an explanation for why prior identification is so objectionable, whether in the case of a “kill list” or in that of bills of attainder, it seems reasonable to think the practice inconsistent with just war theory. Killing in war is a statistical process: each soldier places himself at risk for the sake of advances of the collective. He retains, however, a chance of being among those to survive, and thus the process of waging war doesnot require his overt sacrifice, but rather a personal exposure based on the place occupied by an individual in the context of a collective. A targeted individual, however, does not share in the possibility of survival granted to members of the collective. He is singled out, not for exposure to greater risks than comparably situated others, but for elimination. His risks relative to others fighting on his side are not statistical risks, and do not carry the upside of membership in the war-making collective. He is a criminal awaiting execution rather than a belligerent fighting for a cause or for his own defense.

A subsidiary problem is the criteria for identifying someone by name on a “kill” list, in this case the JPEL. As we touched on in the preceding discussion of the bootstrapping problem, it is not clear it is legitimate to place people on the target list we do not regard as full-blown belligerents in all respects. To elaborate on this point, the issue was raised in a prominent way in the case of Al-Aulaqi v. Obama, Gates and Panetta.43 Al-Awlaki was threatening to U.S. National Security by being a member of a terrorist organization that has hostile intentions towards the United States, and by his acts of propaganda supporting the operation and expansion of that terrorist organization.44 With group membership as the primary criterion for his inclusion in the government “kill list,” we have no assurance that al-Awlaki's inclusion on the list was based on either his status as a belligerent in the traditional sense, or that it was based on a personal self-defense claim, that would also suffice to bring it within the permissible grounds for fighting in war. The threat he posed might not have risen to the level of full-blown belligerency, and it might also have failed as providing a basis for the exercise of self-defense against him by U.S. agents, due to lack of imminence. In this case al-Awlaki's representation on the targeted list arguably should not have been justified in terms of the privilege to kill in war, and his status as a combatant depends on an attenuation of the laws and principles of reciprocal war. Without belligerency on the one hand, or an imminent threat on the other, that might justify killing him (p.172) in the name of national self-defense, the targeting of al-Awlaki was not defensible from the standpoint of the traditional theory of fair combat.

IV. Using drones to kill

A distinctive, though by no means necessary, feature of targeted killing is that it often takes place by drone technology.45 Are there any morally relevant differences between killing with a remotely controlled drone and killing in hand-to-hand combat? As mentioned above, one important difference between “manual” killing by visually sited gunfire and electro-optically sited “distance” killing is that the “target” is more likely to be deprived of the opportunity to surrender. One should, however, be more precise: where drones are used, the target is likely to be deprived of the opportunity to surrender at the point at which surrenders typically occur in traditional combat, namely death. This also has implications for the duty to capture rather than kill: assuming the targeting takes place at a distance, and the missile is fired in an autonomous “fire and forget” mode,46 the target could not, even in theory, effectuate a surrender, since one cannot surrender to a drone. This in turn substantially enhances the difficulty for the killer of fulfilling his obligation to capture rather than kill. In this regard, the use of drone technology locks the aggressor into a killing scenario, where the initiation of the plan of action starting with the placing of an individual's name on the kill list, and committing to effectuate that killing with a technology that enables killing at a distance.

One is tempted here to draw a parallel to the domestic criminal law cases involving spring guns, in which the firing device is automatically triggered by the presence of an intruder on the protected property. Courts have found the use of automatic firing devices of this sort to be impermissible, on the ground that they constitute a commitment to a course of action that is both irreversible once initiated and highly subject to error.47–48 The use of the spring gun is thus reminiscent of the rational choice strategy of binding oneself to the mast, or precommitment. Such automatic plan execution devices require an agent to weigh the costs and benefits of an entire course of conduct up front, in an ex ante position of choice. The difficulties courts have found with spring guns underscore the general objections to (p.173) any kind of automatic plan execution: while the automaticity allows the planner to execute a course of action that might otherwise have been rationally or morally foreclosed to him, the benefits from automatic plan execution are often quickly outweighed by the inability to reassess costs.

In drone killings, however, the killing itself is not “automatic” in the sense identified in the spring gun cases. The drone is manually operated from a remote location, and missiles launched can be under continuous guidance and observation of the target. Still, even in the face of continuous guidance, the likelihood of deflecting the killing into a non-lethal course of action is greatly reduced as compared with manual, person-to-person killing. As the court said in Katko v. Briney,

A possessor of land cannot do indirectly and by a mechanical device that which, were he present, he could not do immediately and in person. Therefore, he cannot gain a privilege to install, for the purpose of protecting his land from intrusions harmless to the lives and limbs of the occupiers or users of it, a mechanical device whose only purpose is to inflict death or serious harm upon such as may intrude, by giving notice of his intention to inflict, by mechanical means and indirectly, harm which he could not, even after request, inflict directly were he present.49

The court's thought here is that because there is no privilege to protect one's home with deadly force against a non-violent trespasser, the same cannot be accomplished by means of an indiscriminate mechanical device that “automatically” targets anyone who enters the property, whether malevolent or excused, justified or merely misguided. The same point can be made about drones: the element of human judgment is eliminated when distant technologies are used to implement decisions about life and death, and this plays an essential role in justifying the decision to kill. Though technically reversible, decisions to target subjects with remote technology obviate the role of human judgment that would most readily allow for reversal or adjustment based on the target's demeanor or activity. As such it operates like a mandatory death sentence that leaves no room for individual mitigating evidence.50

Finally, even if one were to regard being “locked in” to killing rather than capturing as desirable, and hence justified from the standpoint of military necessity, there is a question whether it is morally acceptable to conduct killings of suspected, low-level terrorists at a great remove, given that the distance between attacker and victim also minimizes the opportunity for human intuition and appropriate empathy to play a role.51 Consider the heart-wrenching scene in The Mascot, a memoir written by a son of the tragic tale of his survivor father, who fled the small village of Koidanov, Belarus at the age of five after witnessing his mother and two (p.174) siblings, along with several hundred other Jews from their village, murdered and thrown into a pit by a Belarussian Einsatzgruppen.52 After wandering through the forest for a period of months, he was caught by a Latvian SS unit in the process of executing a group of Jewish prisoners who were at that moment lined up against a Church wall waiting to be shot. For some reason, the boy at that moment asked the Commander, who was about the pull the trigger, if he could have a piece of bread. As the author explains, the simple humanity of the request, the reminder that the executioner and victim were both embodied and creatures of appetite, led the Latvian SS officer to remove the boy from the line-up.53 The shared humanity of the moment of rescue depended upon the physical proximity of attacker and victim—the interlocking gazes and the officer's consequent ability to see a piece of himself in the desperate, ill-fated child. The more distance, the less interaction; the less interaction, the weaker the tug of humanity, that can, on occasion, lead to spontaneous acts of mercy.

For all of the above reasons, I conclude that the traditional justification for targeted killing in war fails with respect to the attenuated cases of modern warfare. Failing the just war hypothesis about why such killings might be justified, there is a second possibility that merits consideration, namely that an alternative reason for killing in war is that there is an imminent threat to the personal security of combatants. While the just war and the personal defense rationales for killing in war operate substantially differently, both are valid in traditional combat, and both provide a justification for targeted killing in the context of traditional battlefield warfare. If the status-centered approach of traditional just war theory fails to justify a practice of advance identification of quasi-combatants, could the realm of agent-relative, personal justifications fare any better?

V. Belligerency, law-enforcement and self-defense

Thus far I have argued that there are at least two reasons why traditional just war theory does not readily accommodate the practice of targeted killing. Although we have not explored the traditional justification for killing in war in any depth, the relevant normative framework is that provided by the concept of belligerency. It is (p.175) the notion of belligerency that makes it permissible for one soldier to kill another in war, and for him to kill not just defensively, namely because he is attacked, but offensively as well, namely in pursuit of victory for his side in the battle, regardless of the danger posed by the enemy soldier. The concept of belligerency is also what makes it permissible for one soldier to kill an enemy soldier regardless of the moral status of their country's claim to justice in the cause of war. The status of a solider as combatant in war carries with it the justification to kill and be killed, because the relationship among enemy combatants reflects a “moral equality” in which each combatant, regardless of the moral status of his country's cause, possesses an equal entitlement to kill soldiers of the other side in virtue of his own adherence to a set of neutral rules governing the waging of war.54 These are the jus in bello rules for conduct war: those that govern how war is waged, rather than the jus ad bellum, namely why it was waged in the first place. The “moral equality of soldiers,” as Walzer puts it, is a function of adherence to the jus in bello, rather than the jus ad bellum, or the justness of an enemy's cause in war,55 and belligerency is a concept that operates in the jus in bello, rather than in the jus ad bellum. The right to kill the enemy soldier at all times, then, is a status-dependent justification for killing, rather than an act-or character-dependent criterion.

Jeff McMahan, by contrast, rejects the traditional thesis of the moral equality of soldiers, but he nevertheless subscribes to the thought that there is a justification for killing in war that stands apart from the justification for killing in other situations. For McMahan, it is the combatant whose cause reflects the true jus ad bellum, meaning that his fight reflects the just cause of his country in going to war. The individual combatant inherits his country's right to wage war on the side of justice, just as the individual soldier fighting an unjust war is deprived of any such right, and the fact that his conduct adheres to the jus in bello does not immunize him from moral criticism based on his inherited reason for fighting.56

On either account—Walzer's or McMahan's—the justification for killing in war remains particular to the domain of war, despite the imperfect parallels that both authors attempt to make between warfare and the concept of self-defense in domestic criminal law. The distinctiveness of just war theory is particularly clear in Walzer's account. The moral equality thesis would appear to be unique to war, and the attempt to find parallels in the personal morality of life-and-death situations seems to elude the Walzerian theorist.57 McMahan, by contrast, is explicit both in rejecting the moral equality thesis, and in claiming that the moral significance of the jus ad bellum over the (p.176) jus in bello finds a parallel in the domestic rules regarding justification, in particular with regard to the law of self-defense. The rules of combat are, by McMahan's lights, mirrored in the normative relationship of agents aggressing against one another in ordinary morality, and outside the context of war. There too, McMahan suggests that the strength of an agent's right to kill another person depends significantly on the moral justifiability of the aggressor's cause as compared with that of the self-defender's.58

Yet, by McMahan's own admission, various domains of domestic law, largely those areas that we think of as highly reflective of ordinary morality, would require significant revision if we attempted to maintain the parallels between the ad bellum/in bello split of traditional just war theory in the context of ordinary morality. McMahan believes that although the law of self-defense, and its corresponding principles in ordinary morality, provide a fairly compelling match for just war theory, the law of self-defense would require considerable revision to fully capture the moral logic that undergirds it. To cite just one example, McMahan thinks that an individual attacked who is lacking in an overall justification for killing based on the superior morality of his cause, and who is confronted with a morally adequate basis for killing him on the part of his adversary, should, strictly speaking, just stand there and allow himself to be killed.59 This is because he has no moral ground for attacking the adversary who is fighting a just war, as the fact that he is being attacked is not strictly speaking relevant to answering the question whether he has a justification for killing a belligerent attacker. But since it is not reasonable to expect a person attacked by an aggressor, even a justified aggressor, to remain still and allow himself to be attacked without counter-response, McMahan is willing to say that the non-justified victim of an attack, although not justified, is nevertheless excused if he exercises self-defense in the face of a justified attack.60 Thus although the law of war and the domestic law of self-defense may on the deepest normative level display the same logic of moral justification, the law of self-defense as currently structured fails to reflect the law of war, and hence in practice, though perhaps not in theory, the law of war and the domestic law of violence come apart. On either account, the right to kill in war is conceived of as isolated from the rest of the law on the use of force in our personal morality and in our domestic legal provisions.

Perhaps the most significant divergence between ordinary morality and just war theory is that the right to kill in the latter context relies on a status justification—that of belligerency—in a way that is never the case in domestic criminal law. (p.177) The aggressor in a case of self-defense is the fitting object of violence only if, and to the extent that, he is aggressing against the defendant. Once his aggression is discontinued, the legitimacy of attacking him also ceases, and he regains his right not to be killed. The same is true in law enforcement: a person who poses a danger to others can be targeted and killed by police, but only insofar as he continues to pose a danger, and as long as he could not be apprehended by non-violent means. The same holds true of defense of others: one is privileged to defend third parties with force in certain instances, but only when, and to the extent, that the first party would have the right to defend himself under the circumstances. What is noteworthy about standard criminal law justifications in this context is that although the right to kill in just war theory is considerably broader than the right to self-defense and other standard criminal law justifications, the latter set of reasons to kill provides in another respect a more expansive entitlement than just war theory.

For example, the right to kill in war is limited to combatants, even if on standard accounts at least, it is an “expansive” right in the sense that even combatants whose cause is lacking in moral justice can avail themselves of the entitlement. For this reason, it is a violation of the laws and morality of war to kill an enemy combatant if the soldier is aware of a substantial risk he will kill a non-combatant civilian in the process. But, for example, on the theory of self-defense this concern is less pressing. If a terrorist threatens to throw a bomb at you, and the terrorist is standing next to his wife, who is unaware of his aggressive plot to deprive you of your life, you are entitled to throw a bomb preemptively at him in order to save your life, despite the fact that you will almost certainly kill the innocent wife in the process. Killing in war, therefore, may not be as constrained with respect to the lives of bystanders as it is under the parameters of just war theory.

This is a curious result for the right to kill in war. It is perhaps for this reason that commentators on the laws of war see the private right to self-defense of soldiers as absorbed into the rights of enemy combatants.61 To allow such defenses to exist side-by-side appears to threaten the coherence of the laws of war, since as “combatant” the soldier may not target enemy combatants where to do so would threaten the life of a bystander. If the justification for killing were based on an individual's private right to self-defense, or on the privilege to protect third parties, the application of the theory (p.178) would be substantially broader. If the right to engage in targeted killing were somehow predicated on one of the criminal law justifications, this would substantially reduce the need for the combatant to concern himself with collateral damage.

If an individual satisfies the criteria for belligerency, he is no longer treated as an attenuated threat to national security; he is viewed as an immediate threat, and he is then a legitimate target by virtue of his status, provided that the killing is done in a way that respects other independent boundaries, such as the restriction on causing disproportionate collateral damage. Self-defense or other-defense, by contrast, knows no limitations based on the identity of the person threatened, and therefore it does not matter whether the use of force is initiated by military personnel, CIA agents, or others. But, I shall argue, when an individual poses a threat to national security interests, but that threat is non-imminent,62 or it is not immediately necessary to respond with force, and when the person of the responder is not personally threatened with an imminent use of force (or that it is immediately necessary to use force by way of response), the permission to use targeted killing fails.

In the previous section we saw that killing in war according to the belligerency rationale may fail to justify the practice of targeted killing, and this is for several reasons. First, the practice pre-names individuals to be assassinated, a practice that goes significantly beyond the military practice of killing in battle and the level of “targeting” it involves. Second, as actually practiced, targeted killing does not restrict the “kill list” to those who would traditionally be considered belligerents. Finally, I raised concerns about the usual way in which targeted killing is conducted, namely with drone technology. While this does not go directly to the justification for the practice, insofar as one can engage in targeted killing without drones, it does raise some doubts about the current form the practice typically takes, as well as raising some basis for reflecting back on the rationale for targeted killing in traditional just war theory. On the other hand, the criminal law justifications are highly likely to fail with respect to those on a named kill list, because the threat they pose is clearly going to be regarded as non-imminent. If just war theory and self-defense both fail to explain the legitimacy of killing according to a named target list, is there any other justification for the practice that would help to justify it?

VI. Preemptive killing in war

In this section, I shall argue that there is a justification for the use of force that falls into a middle-tier category between the status-dependent law of war and what would (p.179) amount to an extrapolation of personal defense in criminal law, as applied to a targeted killing situation. The kind of justification I have in mind is most clearly demonstrated by certain domestic law enforcement circumstances. I shall refer to such cases as instances of “preemptive” force. In these cases, I argue, despite the fact that neither of the standard justifications of killing the enemy combatant under the laws of war, or self-defense against an imminent or immediate threat of serious bodily injury or death is applicable, I shall claim that the use of targeted killing as an instance of preemptive force renders the practice justifiable, subject to certain more restrictive conditions having to do with the apprehension of suspects and avoidance of collateral damage. Thus the concept of “preemptive force” may explain why it is sometimes permissible to kill nonbelligerents who are not posing an imminent threat to one's own or another's security. At the same time, however, it is crucial to understand that the entitlement to kill that falls under this rationale will be of a more limited sort than the justifications that stem from the entitlements that flank it—the killing of belligerents in an armed conflict, justified under traditional laws of war, on the one hand, and the killing of those posing an imminent threat to one's person or the person of another, justified by the traditional self-defense defense, on the other.

I shall approach the topic of preemptive force in war by focusing at first on a closely related topic, the topic of preventive force. Preventive force is primarily force exercised in self-defense or defense of others. It will be helpful to approach the topic from this direction, since there has been a great deal of confusion in recent years about the limits of preventive force. Since that concept is very often equated with self-defense (a mistake), it will be important to explain the privilege to kill in prevention, and to distinguish it from the more specific and slightly different concept of self-defense. We will then return to the laws of war to compare the right to preemptive force with the offensive entitlement to kill fellow belligerents in an appropriately identified armed conflict.

It is sometimes permissible to use more force to prevent harm than it is to punish instances of that same harm. In many cases, for example, the victim of an assault may use deadly force to defend herself against a harm that could not be permissibly punished with death. It is even permissible in many jurisdictions to use deadly force in defense of habitation, but no jurisdiction has ever authorized the death penalty for intrusions into one's home alone. Finally, the police may use lethal force to pursue a fleeing suspect who is resisting arrest, as long as they suspect him of having committed a felony, believe he poses a risk of future felonious activity, and have warned him of their intent to use force if he does not submit to custody.63 In most such cases, however, the felony for which the use of lethal force is authorized is not murder, and hence would not merit the death penalty as punishment. These examples permit a generalization about the relation between preventive and (p.180) retributive force, namely that the extent of permissible preventive force is broader than the extent of permissible retributive force. Thus if targeted killing is conceived as a form of punishment for prior terroristic activities, or as a method of deterring other terrorists (one of the standard functions of punishment), its scope will be narrower than if it is conceived as a method of preventive law enforcement.

While the foregoing states an oft-made point, scholars of domestic criminal law tend neither to opine about the source of this curious dichotomy, nor to trace its implications for other aspects of the criminal law. Discovering the philosophical foundation of a doctrine will often shed light on the scope of its implications. While I cannot undertake a full analysis of preventive force and its relation to retributive force in the current context, I do wish to focus on one particularly helpful aspect of this dichotomy. Suppose one subscribes to a deterrence theory of punishment, according to which one seeks both to explain the current structure of criminal sanctions and prescribe modifications in that structure according to a theory of adequate general deterrence. Then the preventive privilege cannot be explained in deterrence terms as well. The reason should be clear: if the amount of punishment generally prescribed by way of retribution is thought roughly adequate for deterring rational prospective criminals, relative to a desired baseline reduction in the societal levels of crime, then the amount of force authorized by way of prevention would represent overdeterrence relative to that same baseline. In short, if it is not necessary to use the death penalty to deter assault, then the preventive permission to kill in order to avoid being assaulted cannot be required for deterrence purposes, and must be explained according to some non-deterrence-based logic.

What sort of logic would that be? The authorization to use preventive force in self or other defense, or even to prevent the commission of a variety of offenses that are not against the person, must be explained according to one of the available alternatives to deterrence theory. Preventive force may be permissible because it involves the exercise of a right, meaning that it stems from some set of deontological norms. Alternatively, it may be permissible because it represents the expression in some sense of our non-negotiable rational agency. Such would be the case on a social contract picture of preventive rights. The implications of the comparably more expansive authority to use preventive force, as compared with retributive punishment, would then point in either of two directions: instances of social prevention would either be explicable as an expression of a set of societal deontological rights, or as an implicit global covenant for self-protection based on mutual advantage. I tend to the latter view. That view is both sensitive to the interest all have in deterring violent activities and the need to respect basic autonomy conditions in the face of the right to kill in war. The traditional laws of war contain, at their core, a set of rational restrictions on the offensive right to kill. In the framework I am advancing, these limitations can best be understood as premised on mutual advantage. When countries observe the (p.181) restrictions on aggression contained in the laws of war, they fare better than those countries that do not. This assertion, however, only holds as long as the forbearance from unbridled aggression in a conflict with the enemy is mutually observed. Where it is not mutually observed, war degenerates into “total war,” a Hobbesian state of nature in which “every man is enemy to every man.”64 Because Al Qaeda, and groups with terroristic aims, can be expected to violate these conditions of mutuality, it is not possible to apply the rules of reciprocal engagement to collectivities of this sort. It is this insight that provides the most support for the thought that we must treat those who would plunge us in to “asymmetrical” warfare as agents in a self-produced condition of “total war.” But there may as yet be obligations to bring such individuals into the fold of reciprocity, by continuing to respect, albeit unilaterally, the basic conditions of civilized, humanitarian discourse.

In keeping with the foregoing observations from domestic criminal law, I shall elaborate a distinction between two kinds of preventive killing. The first I shall call simply “preventive” killing, and the second I shall refer to as “preemptive” killing. Preventive killing, when justified, is dependent on the need to physically put a stop to the use of force on the part of another. Preemptive killing, by contrast, bears a more attenuated relation to the harm it is designed to forestall; its permissibility follows from the use of rational techniques legitimately employed to dissuade a potential aggressor from following through with his course of action. In particular, it often depends on the fact that it is sometimes permissible to threaten to inflict a harm in order to deter another from the use of violence. In such a case, when deterrence fails and when the threat constituted a legitimate response to the fear of force on the part of another, it may be permissible to follow through on a threat it was morally permissible to issue, despite the fact that the threatened action would not have been permissible as straightforward preventive action. Such is arguably true of law-enforcement action, despite the fact that law enforcement is thought to preclude military enforcement, and that these two types of enforcement are fundamentally at odds with one another. The concept of preemptive killing provides a sound model of the crossover between law enforcement and military. It also provides a useful example of targeted killing, or at least a model that situates the practice correctly relative to its own ambitions, and allows it to be normatively justified as well.

My suggestion is that targeted killing is permissible when it falls squarely into the category of justified preemptive killing. It is rarely, if ever, justified as a form of preventive action. This places certain constraints on the legitimate reach of targeted killing that would not apply if the action must be considered purely preventive. But these restrictions do not seem overly stringent when articulated in the context of the practice of targeted killing. They imply, for example, that targeting must be preceded by a threat to use force, along with an attempt to apprehend the source of the threat. This squares (p.182) with the duty to capture rather than kill that applies to attacks on non-combatants, along with the requirement that applies to everyone, namely that individuals have a right to surrender. Respecting these deep sources of respect and reciprocity in war, despite the fact that one's enemy does not, may be a moral duty, just as the obligation of the police to extend and protect certain rights of suspects cannot be derogated from on the grounds that the suspects themselves do not observe the same conditions with respect to us. Considering targeted killing a form of preemptive killing also has ramifications for the crucial questions of who may be targeted, the extent to which bystanders may be endangered, and other policy aspects of the current debate over targeted killing. It is beyond the scope of this chapter to trace all the implications of adopting a preemptive, rather than a preventive, framework for analyzing targeted killing. The rough outlines of those implications can be easily identified, however: While a preventive framework uses force to impede infringements of self or societal security in a given case, a preemptive framework accomplishes its objective by implicitly or explicitly threatening the use of force to deter such infringements in a given case. Preemption is thus like prevention in that it is particular to a given case, but it partakes in the indirect logic of rational deterrence to induce compliance. Unlike prevention, it appeals to the reason of the agent on whom is operates. It alters behavior by changing the payoffs for (roughly speaking) rational agents, and thus appeals to their ability to project their reasons for acting into the future.

By linking the logic of targeted killing to preemption, I am suggesting that the legitimacy of this technique lies not in the relabeling of the target as a kind of “combatant,” such that it becomes legitimate to target him without the implicit duty of capture. Instead, I am arguing that the permission to target between non-co-belligerents is linked to the legitimacy of issuing a threat to use force if the other does not surrender. The threat is important, because it is an enactment of the duty to capture, in recognition of the fact that one is operating in a non-traditional combatant situation. When the threat is ignored, and the duty to capture cannot be met in other ways, it is then permissible in some cases to follow through on the threat. This places the legitimacy of targeted killing outside the traditional co-belligerency setting, firmly within the purview of practices like nuclear deterrence: it is permissible in some cases to threaten to use force that it would not be, by itself, permissible to use to deter wrongful action. And when the threat is ignored, it becomes permissible to carry through on the action threatened because the demand implicit in the threat was legitimate and the target had the opportunity to conform his behavior to his best reasons for acting by presenting himself for surrender. If this is correct, we can say that the central difficulty with the practice of targeted killing is that it ignores the importance of the public warning or threat. The target is entitled to the opportunity to surrender, and the targeting is legitimate only if the target has been fully afforded the possibility of conformity to the terms of his obligations as expressed in the demand for surrender.


(1) President Obama issued an executive order authorizing the continuation of a system of permanent detention for terror suspects detained in the course of fighting the war on terror. Periodic Review of Individuals Detained at Guantanamo Bay Naval Station Pursuant to the Authorization for Use of Military Force, Exec. Order No. 13567 (March 7, 2011). Criticism of this Order was immediate and unambiguous. The Center for Constitutional Rights (CCR) said: “Today's executive order . . . codif[ies the lawless] status quo. The creation of a review process that will take up to a year is a tacit acknowledgment that the Obama administration intends to leave Guantanamo as a scheme for unlawful detention without charge and trial for future presidents to clean up.” Press Release, Center for Constitutional Rights, “CCR Condemns President Obama's Lifting of Stay in Military Tribunals” (March 7, 2011), available at 〈http://ccrjustice.org/newsroom/press-releases/ccr-condemns-president-obama%E2%80%99s-lifting-of-stay-military-tribunals〉. Executive Director of the ACLU, Anthony Romero, issued the following statement: “providing more process to Guantánamo detainees is just window dressing for the reality that today's executive order institutionalizes indefinite detention, which is unlawful, unwise and un-American.” Press Release, ACLU, “President Obama Issues Executive Order Institutionalizing Indefinite Detention” (March 7, 2011), available at 〈http://www.aclu.org/national-security/president-obama-issues-executive-order-institutionalizing-indefinite-detention〉 accessed November 4, 2011.

(2) See Harold Koh, Legal Adviser to Dept. of State, Address at American Society of International Law, “International Law and the Obama Administration” (March 25, 2010), available at 〈http://www.state.gov/s/l/releases/remarks/139119.htm〉 accessed November 4, 2011 (defending the use of targeted killing as part of armed conflict with enemies in the war on terror).

(3) Tara Mckelvey, “Inside the Killing Machine,” Newsweek, February 13, 2011.

(4–5) There were 1,172 reported kill or capture raids completed between March 2011 and September; see Alex Strick van Linschoten and Felix Kuehn, “A Knock on the Door: 22 Months of ISAF Press Releases” p. 12, available at (https://www.afghanistan-analysts.net/uploads/AAN_2011_ISAFPressReleases.pdf).

(6) Kenneth Anderson, Targeted Killing in U.S. Counterterrorism Strategy and Law 7 (Series on Counterterrorism and American Statutory Law, Working Paper, May 11, 2009), 〈http://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID1415070_code235051.pdf?abstractid=1415070&mirid=1〉 accessed November 4, 2011.

(7) See U.N. Gen. Assem. [GAOR], Hum. Rts Council, Report of the Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions (May, 2010) (prepared by Philip Alston) (criticizing use of targeted killing in areas beyond zone of hostilities); Mary Ellen O'Connell, Unlawful Killing with Combat Drones: A Case Study of Pakistan, 2004–2009 (Notre Dame Law Sch. Legal Stud., Research Paper No. 09-43, 2010), 〈http://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID1654055_code1212987.pdf?abstractid=1501144&mirid=1〉 accessed November 4, 2011.

(8) See Scott Wilson and Al Kamen, “‘Global War on Terror’ Is Given New Name,” The Washington Post, March 25, 2011, available at 〈http://www.washingtonpost.com/wp-dyn/content/article/2009/03/24/AR2009032402818.html〉 last accessed November 2, 2011. In a memo purportedly at the direction of the Office of Management and Budget, the Defense Department's office of security review noted that “this administration prefers to avoid using the term ‘Long War’ or ‘Global War on Terror’ [GWOT.] Please use ‘Overseas Contingency Operation.’”

(9) J.S. Bybee, “Memorandum for John Rizzo: Acting General Counsel of the Central Intelligence Agency” memo, US Deparment of Justice, August 1, 2002.

(10) See, e.g., Convention (IV) relative to the Protection of Civilian Persons in Time of War. Geneva, 12 August 1949.

(11) 18 U.S.C. § 2340A.

(12) “Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment,” General Assembly resolution 39/46 of 10 December 1984.

(13) See Al-Aulaqi v. Obama, No. 10–1469 (JDB) (D.D.C. December 7, 2010). For more information on the lawfulness of operations against propagandists, see Lawrence Preuss, “International Responsibility for Hostile Propaganda against Foreign States,” The American Journal of International Law, Vol. 28, No. 4 (Oct., 1934), 649–68; and Frits Kalshoven, Reflections on the Law of War: Collected Essays (Leiden, Martinus Nijhoff Publishers, 2007), 493.

(14) Gary Solis also makes the point that this potentially makes these officials legitimate targets themselves. See “CIA Drone Attacks Produce America's Own Unlawful Combatants,” Wash. Post, March 12, 2010 (“It makes no difference that CIA civilians are employed by, or in the service of, the U.S. government or its armed forces. They are civilians; they wear no distinguishing uniform or sign, and if they input target data or pilot armed drones in the combat zone, they directly participate in hostilities—which means they may be lawfully targeted.”).

(15) For an argument against this position, see Richard V. Meyer, “The Privilege of Belligerency and Formal Declarations of War,” in this Volume, ch. 7. Responding to Claire Finkelstein, “Responsibility for Acts of War” (2011) (unpublished manuscript, on file with author).

(16) Remarks presented in the Frontline program Kill/Capture by John Nagl, a former counterinsurgency adviser to the former commander of forces in Afghanistan, and General Petraeus, current Director of the CIA. Discussed in Kevin Govern, “Operation Neptune Spear: Was Killing Bin Laden a Legitimate Military objective?,” in this Volume, ch. 13.

(17) Walzer, Just and Unjust Wars, 4th edn (Basic Books, 2006)

(18) Michael L. Gross, Moral Dilemmas of Modern War: Torture, Assassination, and Blackmail in an Age of Asymmetric Conflict (Cambridge University Press, 2010) 102ch. 7

(19) The Model Penal Code, § 2.10 provides: “It is an affirmative defense that the actor, in engaging in the conduct charged to constitute an offense, does no more than execute an order of his superior in the armed services that he does not know to be unlawful.”

(20) “Threats and Preemptive Practices,” 5 Leg. Theory 311 (1999)

(21) SeeMichael L. Gross, “Assassination and Targeted Killing: Law Enforcement, Execution or Self-Defense?” 23 J. Applied Phil. 323, 324 (2006)

(22) See, e.g., Nigel West, “International Special Operations Forces 2010 Year in Review,” June 15, 2011, available at 〈http://www.defensemedianetwork.com/stories/international-special-operations-forces-2010-year-in-review/〉 last accessed November 2, 2011. See also Philip Alston, “The CIA and Targeted Killings Beyond Borders,” September 2011, at 43–4, available at 〈http://ssrn.com/abstract=1928963〉 last accessed November 2, 2011.

(23) See, e.g., Al-Aulaqi v. Obama, supra n. 13.

(24) McKelvey, supra n. 3.

(25) Solis, supra n. 14.

(26) See Hague Convention (IV) Respecting the Laws and Customs of War on Land, Annex art. 23 (b) (1907) (defining qualifications of “belligerents” as requiring “a fixed distinctive emblem recognizable at a distance”). See also Protocol Additional to the Geneva Conventions of August 12, 1949 and Relating to the Protection of Victims of International Armed Conflicts (Protocol I), June 8, 1977, 1125 U.N.T.S. 3, art. 37 (1)(c) (“It is prohibited to kill, injure or capture an adversary by resort to perfidy [including]. . . . the feigning of civilian, non-combatant status.”). For a criticism of this view that uniforms matter, see Christopher Kutz, “The Difference Uniforms Make: Collective Violence in Criminal Law and War,” 33 Phil. & Public Aff. 148, 180 (2005). For more information on recent developments in this view, see Kevin H. Govern, “The Hunt for Bin Laden—Task Force Dagger—On the Ground with the Special Forces in Afganistan: Reviewed by Lieutenant Colonel Kevin H. Govern,” 179 Mil. L. Rev. 210, 217–18.

(27) Anderson supra n. 6.

(28) See O'Connell, supra n. 7.

(29) For one response to this question, see Kevin Govern, “Operation Neptune Spear: Was Killing Bin Laden a Legitimate Military Objective?” in this Volume, ch. 13.

(30) For one response to this question, see Kenneth Anderson, “Efficiency in Bello and ad Bellum: Making the Use of Force Too Easy,” in this Volume ch. 14.

(31) Kevin Govern, “Expedited Justice: Gaddafi's Death and the Rise of Targeted Killings,” JURIST—Forum, Oct. 25, 2011, available at 〈http://jurist.org/forum/2011/10/expedited-justice-the-trend-to-kill-over-capture.php〉 accessed November 2, 2011.

(32) On This Day: 1988: “Jumbo jet crashes onto Lockerbie,” BBC News, December 21, 1988, 〈http://news.bbc.co.uk/onthisday/hi/dates/stories/december/21/newsid_2539000/2539447.stm〉 accessed November 2, 2011.

(33) For instance, as the leader of Libya, Qaddafi became Chairperson of the African Union on February 9, 2009. Press Release, African Union, Leader Muammar Gaddafi Visits AU Headquarters (Feb. 5, 2009).

(34) Kristin Eichensehr, “On the Offensive: Assassination Policy under International Law” 25(3) Harvard Law Review

(35) For more on this issue, see Kevin Govern “Operation Neptune Spear: Was Killing Bin Laden a Legitimate Military Objective?,”in this Volume, ch. 13.

(36) On this doctrine see Claire Finkelstein and Leo Katz, “Contrived Defenses and Deterrent Threats: Two Facets of One Problem,” 5 Ohio. J. Crim. L. 479 (2008); Paul Robinson, “Causing the Conditions of One's Own Defense: A Study of the Limits of Theory in Criminal Law Doctrine,” 71 Va. L. Rev. 1 (1985); also see Leo Katz, “Targeted Killing and the Strategic Use of Self-Defense,” in this Volume, ch. 17.

(37) Model Penal Code § 3.04(2)(b)(i).

(38) MPC § 2.09(2).

(39) The point remains controversial in domestic criminal law as well, as some commentators take the view that creating the conditions of our own defense does not eliminate the entitlement to claim it, since from the standpoint of the aims of the criminal law, the self-defender has done “nothing wrong,” even if his own impermissible scheming placed him in a position in which he was able to claim the defense.

(40) I discuss the general point below in commenting on Jeff McMahan's view of the relation between these two concepts. See infra section III.

(41) SeeGuido Calabresi, The Cost of Accidents: A Legal and Economic Analysis (Yale University Press, 1970)

(42) See, e.g., Ryan P. Alford, “The Rule of Law at the Crossroads: Consequences of Targeted Killing of Citizens,” Utah Law Review (forthcoming).

(43) See Al-Aulaqi v. Obama, supra n. 13.

(44) For more on the circumstances under which he was targeted, see Govern, supra n. 31.

(45) Kenneth Anderson, “Targeted Killing and Drone Warfare: How We Came to Debate Whether There Is a ‘Legal Geography of War’” in Peter Berkowitz (ed.), Future Challenges in National Security and Law (Hoover Institution, Stanford University, forthcoming)

(46) On the importance of the firing mode, see Henry S. Kenyon, “Multipurpose Missile Program Accelerates,” Signal Online, June 2009, available at 〈http://www.afcea.org/signal/articles/templates/Signal_Article_Template.asp?articleid=1962&zoneid=263〉 accessed November 2, 2011.

(47–48) Katko v. Briney, 183 N.W.2d 257 (Iowa 1971)

(50) See Woodson v. N.C., 428 U.S. 280, 303 (1976)

(51) For more on this issue, see Anderson, “Efficiency in Bello and ad Bellum: Make the Use of Force Too Easy?,” supra n. 30.

(52) SeeMark Kurzem, The Mascot: Unraveling the Mystery of My Jewish Father's Nazi Boyhood, (Penguin Publishers, 2002)

(53) After verifying his Jewish identify, and instructing the boy never to let anyone “pull down his pants,” the officer adopted and protected the child, fitting him with a child-size Latvian SS uniform of his own, and allowing him to live as the “mascot” of the SS unit of which he was a part until near the end of the war.

(54) According to Michael Walzer, this is the central concept in the theory of war. Michael Walzer, Just and Unjust Wars, supra n. 17 ch. 1.

(56) Jeff McMahan, Killing in War (Oxford University Press, 2009)

(57) But see my argument in Responsibility for Acts of War (2011) (unpublished manuscript) to the effect that the moral equality thesis exists in domestic criminal law doctrine as well.

(58) McMahan, supra n. 56, at 15–32.

(61) Hugo Grotius, De Iure Belli ac Pacis, G.L. Williams (trans.), (Clarendon Press, 1950), 2:10–11

(62) For more on the imminence requirement, see Russell Christopher, “Imminence in Justified Targeted Killing,” in this Volume, ch. 9.

(63) Tennessee v. Garner, 471 U.S. 1, 722 (1985)

(64) Hobbes, Leviathan, Part I, Chapter XIII, para. 9.