Justified killing in an age of radically asymmetric warfare

Abstract

The growing ability of the US to kill with impunity in war has prompted some to question whether such advantage challenges the moral justifications for inter-combatant violence. This scholarship, however, has yet to properly clarify both the explicit and tacit role of reciprocal risk within this moral determination. A systematic explanation is needed of the function of risk in the right to kill in war. This article engages with the in bello component of the Just War Tradition to determine: first, the role of reciprocal risk in the moral justifications for killing in war; and, second, the extent to which these justifications may be challenged within conditions of radical asymmetry, exemplified today by the unmanned aerial vehicle exclusive violence of the US. The first three sections of this article each review an account of Just War: conventionalism, revisionism and contractarianism. It is argued that the coherence of each of these moral accounts, particularly in terms of the permissiveness of inter-combatant violence, is grounded in an assumption of collective reciprocal risk. Radically asymmetric conditions of battle render ambiguous what would otherwise be a morally unproblematic use of military violence. This article will conclude by highlighting how this challenge manifests in practice, through analysis of the ongoing unmanned aerial vehicle exclusive violence of the US. The radical differentials of physical risk that characterise this violence threaten to erode the capacity of the US to interpret and apply standard judgements of Just War theory against those it targets.

Introduction

Military imbalance is, of course, nothing new. Throughout history, a range of technologies and practices have emerged that have allowed combatants to kill while, at the same time, limiting their exposure to the physical risk of battle. According to a growing number of scholars, however, the military imbalance of today is distinct. At issue is the emergence of radically asymmetric violence — lethal force actualised by one side in the near-complete absence of physical risk, across the full scope of a conflict zone. These same scholars argue that within conditions of radical asymmetry, the risk differential between opposing parties is so overwhelming as to call into question the moral justifications for lethal force (Chamayou, 2015; Enemark, 2013; Ignatieff, 2000: 161; Kaempf, 2014: 79; Kahn, 1999: 201; 2002; Kels, 2012; Sparrow, 2015: 128–129; Zehfuss, 2011). Implicit within this argument is the assumption that reciprocal physical risk between opposing forces is a prerequisite for morally justified battlefield violence. This analysis is overdue, but nevertheless incomplete. Rather than providing a systematic explanation of the role of reciprocal risk, too many authors merely take for granted its status as a fundamental condition in the licence to kill. Additionally, they have yet to sufficiently explain how this challenge functions in practice. What are its implications beyond the world of moral and political theory? This demands further analysis, both of the precise function of reciprocity within each of the distinct and, at times, contradictory positions of the Just War Tradition, and of the theoretical and empirical challenge of radical asymmetry.

This article engages three accounts of Just War: ‘conventionalism’, ‘revisionism’ and ‘contractarianism’. It first clarifies that within each of these accounts, the right of individuals to injure and kill in battle is not contingent on an assumption of personal physical vulnerability, that is, inter-combatant risk. The failure to observe this has left much of the criticism directed against radical asymmetry vulnerable to dismissal as merely the product of an antiquated understanding of war as an equal contest. It is instead argued that the coherence of the existing moral justifications for lethal force is grounded in an assumption of war as a site of reciprocal collective risk between belligerents. While not a formal requirement itself, this collective reciprocity provides the tacit rationale for the permissiveness of the moral standards of war. Collective risk, it is thus claimed, is not merely an empirically descriptive feature of war, but also a key moral one. The erosion of collective reciprocity through radically asymmetric conditions of battle represents a serious and sustained challenge to the coherence of each of these justifications for killing in war. This article then engages with this challenge empirically, within the context of the unmanned aerial vehicle (UAV) exclusive violence of the US. ‘UAV-exclusive’ refers to the ongoing use of UAVs in the near-total absence of military alternatives, including ground-based forces. The radically asymmetric nature of this violence has undermined the ability of the US to interpret and apply morally acceptable limits on its violence. ¹

Conventionalism and the role of reciprocal risk in justified killing

The Just War Tradition is best understood as a rough consensus, rather than a settled theory. Debate and disagreement has always occurred between differing schools of thought regarding various aspects of the doctrine. This continues today, even in relation to the most fundamental questions of Just War, including the conditions by which individuals may be morally liable for lethal targeting. The first of these competing accounts is conventionalism, long regarded as the orthodox position within the Just War Tradition. This section begins by outlining the main theoretical claims of conventionalism, with a particular focus on the role of collective reciprocal risk in its justification for lethal force. It then argues that the coherence of this justification is undermined within conditions of radical asymmetry.

Moral equality and material threat

Arguably, the most influential contemporary proponent of the conventionalist position is Michael Walzer, who outlined its core tenets in his classic work Just and Unjust Wars (Walzer, 2006a). Within its pages, Walzer sought to provide moral and practical answers to some of the most fundamental questions of war, including when and against whom violence may be directed. Walzer was particularly conscious of the extent to which modernising trends in war had depersonalised battle, and, in the process, driven many chivalric practices into obsolescence. In their stead emerged a new military code, grounded in the servitude of the soldier. It is this very servitude, Walzer argues, that provides the grounds for war’s moral coherence. As philosopher and former soldier J. Glenn Gray (1998: 159) articulates: ‘[t]he foe is a human being like yourself, the victim of forces above him over which he has no control’.

From this assumption, Walzer developed and defended the moral equality of combatants. This concept establishes the realm of responsibility for all fighters as equally limited to that of jus in bello — the moral rules that govern battlefield conduct once hostilities have commenced (Walzer, 2006a: 34–37). Accordingly, combatants on both sides of the conflict possess identical privileges and responsibilities — rights that remain independent of the just or unjust status of the conflict’s overall cause. It is this symmetry between enemies, along with the independence of in bello from ad bellum responsibilities, that constitutes Walzer’s ‘war convention’. With soldiers possessing equally valid claims to moral innocence, the basis for the right to kill, and, just as importantly, the liability to be harmed, must be found elsewhere than direct moral culpability.

To justify the killing of combatants in war, Walzer and others draw upon the logic of self-defence. Combatants may be targeted with lethal force on the basis of the mortal threat they pose to others. According to Judith Thomson’s (1991: 300–301) theory of self-defence, this mortal threat equates to a rights violation — the potential target has a basic claim against the aggressor to not pose an imminent threat to their life. Yet, in war, combatants are armed and specially empowered to produce this very outcome, and kill the enemy in most circumstances. They have become, it is presumed, ‘dangerous men’ (Walzer, 2006a: 145). It is this projection of threat that renders the combatant subject to defensive killing (Norman, 1995: 168; Shue, 2004: 111; Walzer, 2006a: 145; Zehfuss, 2011: 555). As Cheney Ryan (2008: 138) explains: ‘soldiers are “mutual-attackers” whose predicament simultaneously renders them “mutual-self-defenders”’.

By this same logic, it is the presumptive absence of reciprocal risk between aggressor and target that provides the core justification within conventionalism for non-combatant immunity. Thomas Nagel (1972) made tacit reference to reciprocal risk in his 1970s work ‘War and massacre’. Writing against the backdrop of US involvement in Vietnam, Nagel examined the factors that distinguish legitimate killing in war from criminal murder. Nagel (1972: 128) defined murder in the context of battle as ‘the deliberate killing of the harmless’. He went on to write that combatants and non-combatants should be delineated ‘on the basis of their immediate threat or harmfulness’ (Nagel, 1972: 140). As Henry Shue (2008: 100) observes: ‘war is not about killing people who are morally liable to be killed; it is about killing people who may otherwise kill you’. To intentionally kill non-combatants is to kill beyond the scope of self-defence, and thus in violation of an individual’s right to life (Fullinwider, 1975: 94).

The justification for killing combatants — and sparing non-combatants — hinges on the threat represented by the former. It is the material threat posed by the combatant that forfeits their right to life and the reciprocal risk between combatants that ensures an equal right to kill. This is important to reiterate going forward. Conventionalist scholars do not generally regard reciprocal risk in war as a desired end in itself. Reciprocal risk is better understood as the by-product of the desired end: the restriction of violence to the harmful.

This is relevant when determining the extent to which military imbalance, and specifically radically asymmetric violence, challenges the coherence of the conventionalist right to kill. For a moral challenge to emerge in this context, the asymmetry in question must pass a substantially higher threshold than the mere mitigation of inter-combatant risk. It must instead undermine the potential target’s capacity for threat. As the next section outlines, this capacity is sufficiently undermined only when asymmetry impacts war at the collective level.

Collective risk, status and the challenge of radical asymmetry

For conventionalist scholars, the material threat projected by combatants in war is fundamental to the invalidation of their right to life. What is less immediately clear, however, is how imminent this threat must be to render such individuals morally liable to defensive harm. Influential jurist Hugo Grotius (2012: 390) argued that the soldier must ‘spare the blood of his foes … [and] condemn no one to death, unless to save himself from death’. For killing to be justified, Grotius added, ‘the danger … must be immediate and imminent in point of time’ (Grotius, 2012: 83). ² Samuel von Pufendorf (1729: 837) similarly argued that it is ‘lawful for me to use violence against my enemy [only] till I have repulsed the danger he threaten’d me’. Both statements advance a conception of imminence narrow enough to presumably render a significant portion of combatants temporarily or even permanently immune to lethal targeting. Many would rightly consider such a reading as over-stringent. In order to maintain the logic of self-defence while avoiding such unworkable practical outcomes, the conventionalist account draws its coherence from the collective reciprocity of risk between belligerents.

As Noam Zohar (1993: 615) explains, within war, ‘it is a collective that defends itself against attack from another collective, rather than simply many individuals protecting their lives in a set of individual confrontations’. Combatants, even those who serve a non-combat function, are trained to deliver violence, armed and possess the right and ability to defend themselves and take part in operations if deemed necessary. They retain, in short, the capability to pose a lethal individual threat. More importantly, though, the combatants of an army also ‘contribute to the achievement of its ends’ and ‘in time of war they pose a unified threat’ (Walzer, 2006b: 4, emphasis added).

What this means in practical terms is that membership of the military (status) is sufficient in most cases to render combatants morally liable to harm, even those who at the time of their targeting are materially non-threatening (conduct) (Benbaji, 2008: 466; Gross, 2006: 329; Meyer, 2012: 187). It is this permissive logic that Walzer (2006a: 138) draws upon when writing: ‘the first principle of the war convention is that, once war has begun, soldiers are subject to attack at any time (unless they are wounded or captured)’. If enemy soldiers are targeted in flight, or while sleeping at their post, it is only their immediate (inter-combatant) threat that has diminished. They remain a sufficiently contributing element of the collective (inter-belligerent) threat of their side. In situations such as these, ‘the “tacit contract” of “kill or be killed” is still intact, even if attenuated’ (Kels, 2014). ³ There are some notable exceptions to this permissive logic. Johnson (2006: 668) writes, for example, that combatants ‘rendered helpless’ are not to be killed, but rather treated humanely and made prisoners of war. In these circumstances, a combatant is re-individualised on the basis of their absolute defencelessness. ⁴

We may rightly question whether this exemption to the targetable status of combatants goes far enough. After all, war is replete with examples of combatants who, when killed, represented neither a specific individual threat, nor contributed to the overall collective threat posed by their military — at least to a meaningfully greater degree than the protected non-combatants of their side. It is crucial to recognise, however, that the moral justification for targeting combatants within the conventionalist account of Just War rests on the presumption, rather than the certainty, of liability. The reality that combatants within a war zone may be subject to harm at any moment justifies the ex ante estimation that their adversary, no matter the circumstance, is both a specific threat and a sufficient contributing element to the collective threat of their side. While an individualist determination of liability would be morally preferable, it cannot be practically adopted within the kill-or-be-killed circumstances of combat. Only in the most extreme and unambiguous cases of combatant defencelessness is this presumption revoked; in all others, combatant liability is collectivised.

The argument here is that this presumption, which manifests as collective, status-based liability, maintains its coherence only for as long as the conflict zone itself can be realistically conceptualised as a site of two-directional violence. Once asymmetry reaches a level sufficient to insulate the military of one side entirely, or near enough, from the risk of injury and death, this is no longer the case. Within such conditions, the capability of the weaker side to pose a material threat has been profoundly curtailed, not merely in terms of individual encounters, but collectively. The scale of this mitigation of threat undermines the ability of the stronger party to draw upon even attenuated conceptions of self-defence in its justification for lethal force. This constitutes a fundamental challenge to the moral coherence of status-based targeting. Recall again the position of Walzer (2006a: 145, emphasis added): ‘We are all immune to start with; our right not to be attacked is a feature of normal human relationships. That right is lost by those who bear arms “effectively” because they pose a danger to other people’.

Conventionalism attaches considerable significance to the assumption that war, and the violence that constitutes it, is fundamentally distinct from the circumstances and violence of normal society. As Statman (2012: 99–100) notes, in the absence of a state of war, collectivised liability is no longer applicable and lethal force must once more be justified on a less permissive, individualised basis. The argument here is that while war may still technically operate within conditions of radical asymmetry, its nature, grounded in a collective reciprocity of risk, has been so transformed as to strain the moral premise of conventionalism. This is not to argue that the combatants of the weaker side are no longer liable to lethal targeting. What radical asymmetry does call into question, though, is the collective ex ante designation of such liability.

Revisionism and the role of reciprocal risk in justified killing

The moral inequality of combatants

As already outlined, Just War is less a theory than a tradition, a synthesis of various schools of thought. Given the diversity between and within these sources, it is perhaps unsurprising that the tradition itself contains a number of enduring conceptual fault lines. One of the most significant of these is the long-standing and ongoing division — roughly split into two positions — regarding how best to allocate moral responsibility on the battlefield. The first is the conventionalist position explored earlier, which affirms the existence of a ‘moral equality’ between combatants in war, regardless of the ad bellum status of either side. Alongside this exists a rival conception of battlefield morality that draws a closer analogy to policing. Within a significant portion of early Christian and natural law sources, war was viewed as a judicial, unilateral instrument, with belligerent rights reserved exclusively for those with the rightful claim to a just cause (Coates, 2008: 176; McMahan, 2008: 19; Reichberg, 2008). It is this interpretation that is re-invoked by revisionist — sometimes termed ‘neoclassical’ — theorists.

According to this position, a presumption of ‘moral equality’ between combatants fails to conform to the standards of normal morality, which regards the justice of an individual’s cause as central to the permissibility of force, including self-defence. Today, this critique is led by Jeff McMahan and others, ⁵ who highlight that in normal society, one does not render oneself liable to defensive attack simply by posing a threat to another as this would expose those engaged in justified self-defence to permissible violence from their culpable attacker (McMahan, 2008: 21). This, McMahan argues, reveals the inherent flaw in Walzer’s ‘war convention’ as a guideline for moral conduct in war. McMahan (2008: 22) proposes an alternative criterion for liability to attack:

A person is morally liable to attack in war by virtue of being morally responsible for a wrong that is sufficiently serious to constitute a just cause for war, or by being morally responsible for an unjust threat in the context of war.

A key feature of revisionism is the belief that the morality of violence in war is analogous with the morality of individual self-defence. It is only in terms of scale that the two differ (McMahan, 2009: 156). Therefore, although there may be different epistemic thresholds for defensive harm, the substantive moral principles are the same for each. A ‘reductive’ account of war ‘reduces the rights and responsibilities of combatants entirely to the rights and responsibilities of individual persons’ (Rodin, 2008: 47). This shift to an individualist prism marks a significant break with the conventionalist understanding of collectivised liability, and, by extension, its justification for targeting on the basis of combatant status. For revisionists, combatant membership is not in itself sufficient to render an individual an appropriate target for lethal force. Non-combatant membership is, by this same logic, insufficient to automatically shield an individual from targeting (Steinhoff, 2013: 187).

For all its other implications, this shift from a material to moral-based account of liability would appear, at least on its surface, to resolve the potential tension between radical asymmetry and the moral justification for killing in war. The elimination of reciprocal risk between enemies, after all, is problematic only for as long as material threat itself is a relevant factor in the right to kill. Seth Lazar (2015: 106) suggests as much, arguing that an individual’s ‘vulnerability and defencelessness, if morally relevant at all … [is] relevant only when the victim is not liable to be killed’.

However, even within revisionism, an assumption of collective reciprocal risk remains an important factor in the moral designation of combatant liability. Just as with conventionalist accounts, therefore, the coherence of revisionism is challenged within conditions of radically asymmetric violence.

Radical asymmetry and the responsibility threshold

Few aspects of the revisionist position have generated as much debate and controversy as its potential to erode the targeting distinction between non-combatants and combatants. Radical asymmetry amplifies this risk by undermining the presumption of liability among low-level combatants.

It is important to first recognise that most Just War revisionists maintain self-defence as a justifying principle for lethal force, but precondition such a claim on a finding of moral responsibility. Theories of moral-based liability hold that it is morally better to target individuals ‘who are responsible for initiating some unjust threat than to kill the person who, at a given point in time, poses the threat, but bears less responsibility for it’ (Statman, 2012: 96, emphasis in original). By this standard, some (unjust) civilians, on account of their moral responsibility for the collective military threat of their side, will forfeit their immunity from direct and deliberate harm. This would bring to an end the conventional and long-standing targeting distinction between combatants and non-combatants in war. McMahan (2008: 22) himself has admitted that a potential consequence of his position is that non-combatants may be attacked if they bear a ‘high degree of responsibility for a wrong that constitutes a just cause for war’. The danger of this approach is that it will essentially open ‘the floodgates to total war’ (Lazar, 2010: 188).

David Rodin, one of the most influential advocates of the revisionist position, explores this in further detail, concluding that McMahan’s ‘asymmetry thesis’ is only half correct. Rodin accepts that unjust combatants have no moral right to direct lethal force against the just. Crucially, though, moral asymmetry does not grant the just combatant additional in bello privileges, such as the right to target enemy non-combatants (Rodin, 2008: 45). Non-combatants may bear considerable responsibility for an unjust war and, by extension, the unjust threats that manifest therein. However, ‘meeting this threshold of fault for an unjust threat [is not in itself] sufficient for liability to defensive force’ (Rodin, 2008: 49). Even if we factor in the moral culpability of non-combatants, defensive force should still be restricted to those ‘whose intervening action is most proximate to the threat’ (Rodin, 2008: 50). This is echoed by a number of other revisionist authors, who argue that only individuals sufficiently ‘causally connected’ to unjust threats may be justifiably targeted in war (Fabre, 2014: 77; Lazar, 2015: 63).

Unjust combatants, at least according to Rodin (2008: 48), ‘clearly meet the minimum threshold of responsibility to be liable for defensive force on liability accounts of self-defence’. For this reason, just combatants may target them without moral censure throughout the course of the war. In contrast to combatants, non-combatants contribute to the unjust threat of their side to a degree that is insufficient for liability to harm. This emphasis on casual connection/proximity demonstrates one way in which revisionist scholars can preserve the critical distinction between protected non-combatants and targetable combatants.

Yet, as authors such as Seth Lazar (2010; 2013: 411) have highlighted, such an approach must confront what he terms the ‘responsibility dilemma’. This refers to the fact that in war there exists a meaningful percentage of combatants who individually will neither cross the threshold of necessary responsibility for an unjust threat, nor represent a material threat at the time of their targeting. Take, for example, the case of the fleeing soldier. A situation not uncommon in war is that of a defending army who, in response to the fire superiority of the enemy, panic, abandon their weapons and break in disarray. According to most revisionist accounts, killing these individuals would present no specific moral challenge. Nor would it if, instead of fleeing, these same individuals were set upon while sleeping. How, one might ask, can causal connection to a threat be established in the case of these sleeping soldiers? If it cannot, how can their direct and deliberate targeting be justified as a necessary measure of defensive force? Recall again that the revisionist position is a ‘reductive’ account of war. What this means is that in contrast to conventionalism, individuals are not liable to harm on the basis of group membership, even if that group is an unjust one (Lazar, 2015: 131–132; McMahan, 2009: 188). Rather, to be targeted, individuals must sufficiently contribute to the collective threat of their side. How, then, do we square this assertion with the realities of war?

It should first be acknowledged that with regard to mid- and high-level members of the military, who it is reasonable to assume have a more direct role in the organisation and execution of threats, establishing a causal connection will be less challenging. In the case of low-level combatants, however, the challenge is acute. Here, McMahan himself provides an effective response. Even within a ‘reductive’ account of war, McMahan (2004a: 76) suggests that a distinction can be drawn between normal society and the battlefield in terms of threat:

War involves threats that consist of activities organized in phases over extended periods of time. A soldier sleeping in invaded territory has already attacked and is engaged in attacking in the same way that I am engaged in writing this essay even while I pause to make a cup of tea.

Within the boundaries of the conventional battlefield, threats may manifest at any place and time. It is a realm of structural danger. As McMahan (2011: 548–549, emphasis added) writes elsewhere, within most war zones, ‘conditions that would prompt [unjust combatants] to kill have a significant probability of arising’. The presumption, therefore, that all opposing unjust combatants — even those at the lowest level — sufficiently contribute to harm, and are thus liable to harm themselves, is a morally defensible one. Importantly, this holds true even in relation to those combatants who when targeted, lack at that precise moment the capacity to injure and kill. As this illustrates, revisionism, even when applying a reductive individualist lens to its determination of liability, allows for a degree of permissiveness in inter-combatant violence that is well beyond what would be acceptable in most non-war scenarios.

This article argues that the coherence of such a presumption founders at the point at which military imbalance undermines the collective reciprocity of risk that binds belligerent forces. Again, this is not to argue that the low-level combatants of the weaker side are now immune from lethal targeting. What is instead suggested is that within conditions of radical asymmetry, the presumption of combatant liability can no longer be credibly sustained. Instead, the moral presumption inverts: low-level combatants of the weaker side are presumptively non-liable, and they remain so unless a sufficient causal connection to the collective threat of their side can be conclusively established. To maintain the more permissive status quo within the context of radical asymmetry is to risk one of two outcomes. The first would be the projection of injury and death to enemy combatants on the weaker side who, as a consequence of a lack of collective reciprocity, are not fully liable to this fate. The second, and potentially more serious, implication would be a lowering of the liability threshold itself to include a greater number of ‘contributing’ unjust non-combatants. Such a move would place considerable strain on existing moral limitations on violence in war.

The differences between conventionalist and revisionist accounts of Just War are significant. Crucially, though, both draw upon an assumption of collective reciprocal risk, albeit to varying degrees, in their justification for killing in war. For this reason, the coherence of each is challenged by radically asymmetric violence. This is also true in relation to a third position of Just War: contractarianism.

Contractarianism and the role of reciprocal risk in justified killing

The contractarian right to kill

As outlined earlier, conventionalism draws a distinction between in bello and ad bellum responsibilities in war. Free from moral culpability for the cause of the conflict, the equal right of combatants to kill is grounded in their shared exposure to injury and death. Violence is justified according to the basic right of self-defence, a justification that is available to all fighters through the imposition of reciprocal risk. Like conventionalism, and in contrast to revisionism, contractarianism separates in bello restraint from ad bellum restrictions. It defends this separation, however, through a more consequentialist logic.

For contractarianism, immunity to lethal targeting is relinquished by combatants in war, not due to the individual threat they pose, but rather because both sides freely submit themselves to a fair and mutually advantageous norm (Benbaji, 2008: 487). The presumption is that states possess a rational interest in preserving their capability to defend against aggression while, at the time, seek to limit the destructiveness of war. This is ensured through a shared compromise: ‘combatants may be attacked with almost no restrictions, while non-combatants may not be (directly) attacked, with almost no exceptions’ (Statman, 2012: 97). This tacit bargain balances the demand of military necessity with the aspiration of humanitarianism. As with the conventionalist position, the status of combatancy is sufficient grounds for targeting. Contractarianism is distinct, however, in that it justifies this permissiveness without resorting to what many would consider an overly broad interpretation of the right to self-defence (Benbaji, 2008: 487; Statman, 2011: 451).

Within this account, the mutual right to kill is the outcome of a ‘contract’ between enemies, one shaped by informal rules and widely shared attitudes, as well as more formal treaty-based positive and customary international law (Benbaji, 2011: 45). Ultimately, though, the coherence of contractarianism is centred on rational insight into best consequences. This logic is outlined in Thomas C. Schelling’s (1980) classic work The Strategy of Conflict, which employs a game-theoretic approach to the subject of war. Within, Schelling explores how best to impose limitations on violence:

To characterise the manoeuvres and action of limited war as a bargaining process is to emphasise that, in addition to the divergence of interest over the variables in dispute, there is a powerful common interest in reaching an outcome that is not enormously destructive of values to both sides. (Schelling, 1980: 6)

The outcome of this centuries-long bargaining process is the framework of jus in bello. This emphasis on military custom does not diminish the moral seriousness of these rules. According to Walzer (2006a: 44–45), a principal strength of the existing war convention is the extent to which it has been internalised by those tasked with its implementation. Even Jeff McMahan (2004b: 731), arguably the most prominent critic of the current liability distinction between combatants and non-combatants, acknowledges the practical importance of the current rule-set, which has been ‘refined, tested, and adapted to the experience of war’.

The formation of contractually binding limits, even those regarded as arbitrary in the deep moral sense, is crucial to restraining the worst excesses of war, including the zero-sum dynamic of total war. The question this then raises is what impact, if any, does the emergence of unprecedented military imbalance have on the basic functionality of this contract? It is argued here that the in bello right to kill advanced by contractarianism is anchored to an assumption of reciprocal collective risk between warring parties. When this is absent, and the collective threat of one side is nullified, the coherence of this justification for killing in war is undermined.

The one-sided bargain of radical asymmetry

In order to highlight the challenge of radical asymmetry to the contractarian position, the importance of collective reciprocity must first be established. It is self-evidently not the case that combatants in war formally waive their right against being lethally targeted by the enemy. Rather, it is assumed that the status of combatancy, as well as the privileges that go with it, ‘constitutes tacit consent to the rules of an adversary institution’ (Applbaum, 1999: 118). Yet, as Benbaji (2011: 45–46) observes, this tacit consent between enemies holds only for as long as these rules can be realistically conceptualised as part of a ‘symmetrical code’, one that is at its heart ‘mutually beneficial and fair’. A recognition of ‘mutual benefit’ rests on the assumption that: ‘The outcome of following a symmetrical code will be better to each relevant party (i.e., individuals and decent states) — in terms of expected benefits and expected protection of rights — than the outcome of following an asymmetrical code’ (Benbaji, 2011: 46).

Klem Ryan (2012) makes a similar point in ‘Fair and unfair wars’. The rules of war, Ryan (2012: 169) argues:

[N]eed to represent fair constraints on both parties. Indeed rules of restraint in war develop and are operable only when participants in war see themselves as engaged in a practice which includes the enemy as a participant and not merely as the focus of hostility.

Reciprocity, as defined in these terms, cannot exist only in the abstract, but must be observable through ‘mutually demonstrable practices’ (Ryan, 2014: 214). It is argued here that few aspects are as critical to the practical demonstration of this reciprocity than a mutual right, and realistic capability, to kill those deemed liable to harm. French philosopher Grégoire Chamayou (2015: 161) endorses this view, arguing that ‘the decriminalisation of warrior homicide presupposes a structure of reciprocity’. This reciprocal relationship, adds Thomas Hurka (2007: 210), hinges on the acceptance by both parties ‘that they may permissibly be killed in the course of war’. Just as with conventionalism and revisionism, mutual threat is a key assumption in the contractarian justification for lethal force. When one side effectively withdraws its military from exposure to threat, this assumption cannot be sustained.

To be clear, this is not an assumption of equality of outcome — meaning an equal distribution of injury and death to all participants. What this account of Just War does assume, however, is an equality of obligation. In practical terms, this translates into an expectation that both parties to a conflict make available at least some portion of their forces for targeting. For most of history, this expectation has required little elaboration on account of the unavoidability of collective reciprocal risk in war. Within certain contemporary military contexts, though, we have witnessed a fundamental shift in the nature of war, defined by the capability of one side to project violence in the sustained absence of reciprocal vulnerability. Within these extreme conditions of unexposed power, the functional premise of the contractarian bargain is challenged. What, asks Grégoire Chamayou (2015: 162), ‘is the worth of a right to kill one another without crime when only one of the two protagonists can still enjoy the real content of the founding permission?’

Just as with other moral accounts, this challenge does not derive from the capability of select individuals in battle to mitigate risk. The contractarian approach translates into a permissive framework of violence, with all combatants subject to harm, not merely those who at the time of their targeting pose an immediate and direct physical threat. Crucially, though, the implicit bargain that underpins this permissiveness retains its coherence only for as long as violence remains ‘a remedy open to the weak as well as the strong’ (Rodin, 2006: 159). When this no longer holds, and violence is monopolised by one side, the in bello contract is destabilised. While this does not necessarily invalidate the right of the stronger party to kill, it does call into question the moral basis for the permissiveness of their violence. To ignore this concern is to risk rendering the Just War Tradition ‘either an ally to the powerful, or obsolete’ (Zehfuss, 2011: 555).

This article has so far engaged with three competing approaches to Just War: conventionalism, revisionism and contractarianism. The justifications for lethal force advanced by each, it has been argued, are challenged by conditions of radically asymmetric violence. To be clear, this challenge is grounded in coherence, rather than compliance. The fact that such violence fails to explicitly violate the rules of war does not resolve this tension, but, rather, points to a deeper one — specifically, the potential of radical asymmetry to destabilise the very premise upon which these moral rules rest. Conformity with the rules of war matters only for as long as the rules themselves remain effective signifiers of appropriate conduct in battle. Radical asymmetry threatens to undo this by undermining the presumptive liability of targeted enemies. What we are then left with is a diminished moral framework, one less capable of delineating the reach and limits of battlefield lethality. This article will now detail the practical consequences of such an outcome in relation to the American use of UAV-exclusive violence. The challenge of radical asymmetry goes beyond the UAV, encompassing a broader penetration of distancing and robotic technology into the landscape of war. At present, however, this mode of violence best illustrates the associated moral tension.

UAV-exclusive violence

The radically asymmetric potential of UAVs

It is first worth clarifying that UAVs are but one of a number of moral and legal controversies to have arisen in the wake of America’s ongoing ‘War on Terror’. From its genesis, the campaign was framed by the Bush administration as a ‘state of exception’: a conflict of such import that the circumvention of particular international norms, including those of Just War, could be rationalised (Ralph, 2013: 1). ⁶ This logic underpinned a range of questionable actions taken by the US during this period, including its designation of captured enemy fighters as ‘unlawful combatants’.

Among the critics of this approach was Barack Obama, who, as president, sought to justify UAV strikes using long-established frameworks of legal and Just War theory. In a 2013 speech, Obama (2013) defended the US UAV campaign as a ‘Just War’, citing a ‘near-certainty standard’ of no civilian deaths. The extent to which this holds true remains the subject of intense debate. Problematically, though, much of this analysis contains an implicit assumption that should non-combatant safety be maintained, and should killing be restricted only to the liable, then the in bello challenge of UAVs would be resolved. What this overlooks is the potential of radical asymmetry to destabilise the very moral frameworks responsible for this designation of liability. The following section addresses this in detail, highlighting the extent to which UAV-exclusive violence challenges the existing justifications for killing the otherwise targetable in battle.

The use of UAVs by the US is divided between two very different types of battlefield. The first are active ‘hot’ battlefields, such as Afghanistan, in which strikes have been conducted by the US military in conjunction with ground forces and other kinds of air power. UAVs have also been deployed in the absence of a significant American ground presence, in regions such as Pakistan, Yemen, Somalia, Libya and the Philippines (Ackerman, 2012; Ahmed and Martin, 2012; DeYoung, 2011; Horton, 2017; Rohde and Khan, 2004). The first of these ‘cold’ UAV strikes took place in Yemen in November 2002, against suspected al-Qaeda leader Qaed Salim Sinan al-Harithi (Johnston and Sanger, 2002). A Time article published days after the strike described it as an ‘assassination’, but speculated that such violence was ‘unlikely to become a norm’ (Karon, 2002). To the contrary, these controversial strikes have become an essential component of the American response to global terrorism. This decision by the US can be explained in large part by the potential of UAVs to impose conditions of radical asymmetry upon the battlefield.

UAV-exclusive violence is qualitatively distinct from previous episodes of battlefield asymmetry, such as military sniping and manned aerial bombing. In each of these prior cases, ‘a vestigial vulnerability has always remained: snipers can be shot or captured, planes can be shot down, and destroyers can be sunk’ (Gusterson, 2014: 196). Even if we consider one of the most extreme contemporary examples of military imbalance — the firing of cruise missiles from US Navy ships in the Mediterranean against targets in Libya, Iraq and Sudan (see Delahunty and Yoo, 2002: 507–512) — there is a proximity between combatant and target that makes possible, however unlikely, a retaliatory response. UAVs, at least in some cases, kill their target while their operator is as geographically removed from the encounter as it is possible to be. The extent of this distancing has advanced those empowered closer than ever before to the ultimate prize in war — riskless killing: ‘By prolonging and radicalising pre-existing tendencies, the armed drone goes to the very limit: for whoever uses such a weapon, it becomes a priori impossible to die as one kills. Warfare, from being possibly asymmetrical, becomes absolutely unilateral’ (Chamayou, 2015: 12–13).

It should be noted that UAVs, particularly slower-moving and non-stealthy models, remain highly susceptible to the ground defences of most major states. ⁷ While this has no bearing on the physical risk assumed by American operators, it does limit their potential application. At present, radically asymmetric violence is achievable only in a narrow context: when UAVs are utilised to the exclusion of significant ground forces against non-state enemies. Within these specific conditions, however, the imbalance is profound. In many cases, those targeted are not only rendered non-threatening at the precise moment of their incapacitation or death, but are stripped — on aggregate — of their very capability to pose a credible physical threat to the military forces of the opposing side. This represents a transformative shift in the nature of war. UAV violence, when undertaken to the exclusion of higher-risk alternatives, challenges the basic assumption of war as a contestation of collective reciprocal risk. This, it is argued, has the potential to destabilise the coherence of the moral justifications for lethal force, specifically in terms of the permissiveness of inter-combatant violence. This will now be illustrated in reference to the Just War accounts of contractarianism, conventionalism and revisionism.

The moral challenge of UAV-exclusive violence

As outlined earlier, the contractarian account of Just War frames the mutual right of combatants to kill as the outcome of an implicit contractual relationship between collective belligerents. Under this ‘contract’, the non-combatants of both sides are preserved from direct and deliberate violence. Combatants, in contrast, are entitled to kill while, at the same time, targetable themselves to lethal force. This contract risks breakdown when the battlefield is restructured under conditions of unilateral violence. Klem Ryan (2012: 46) makes a similar claim: ‘recent technological changes in warfare have altered the practice of war in ways that inherently undermine the conditions needed for conventions to operate as they traditionally have’.

One of the most destabilising of these technological changes has been the emergence of UAV-exclusive violence. There have been a number of innovations in war — early projectile weapons, firearms, the submarine, military sniping and manned aerial bombing — that have reduced the capability of those targeted to effectively retaliate on the battlefield. Yet, in none of these cases was the threat potential of the targeted truly nullified. The substitution of human combatants for machine proxies goes well beyond these historical asymmetries, challenging the very functional premise of the contractarian justification for killing in war.

One obvious counter to this argument is the enormity of the moral crimes perpetrated by many of those targeted by American violence. As Daniel Statman (2012: 106) writes: ‘[a]ccording to … [contractarianism] nothing could be wrong with the use of … [targeted killing] against organisations like al-Qaeda which have no respect for the conventions of warfare’. ⁸ There is no question that the conduct of some of these non-state organisations, particularly their deliberate targeting of civilians, places considerable pressure upon the adversarial bargain of contractarianism. Crucially, though, these crimes do not negate the pressure also applied by the refusal of the US to expose its own forces to the reciprocal risk of the battlefield. Rather, what we are left with is a two-directional strain on the contract: the immunity of non-combatants by the weak, and the vulnerability of combatants by the strong. ‘It is difficult to imagine’, suggests Lt. Col. Douglas Pryer (2013: 19), ‘how anyone could feel that their enemies were justified in waging war against them via remote-controlled machines … if there were no way they could reply in kind … the situation seems fundamentally unfair, unjust, or unreciprocal’. UAV-exclusive violence challenges the assumption of collective reciprocity upon which the coherence of the contractarian account rests. In the absence of this coherence, the permissiveness of American violence loses much of its moral force. This challenge remains equally significant within the conventionalist account of Just War.

Conventionalism grounds its justification for inter-combatant violence on the material threat, rather than moral culpability, of those targeted. Drawing upon the logic of self-defence, this position morally empowers combatants to target those on the battlefield that pose a mortal danger, that is, other combatants. Some have argued that UAVs neutralise the physical risk of their operators to a degree that calls into question the coherence of the conventionalist justification for lethal force. According to Maja Zehfuss (2011: 555):

The use of weapons that put combatants out of their enemies’ range invalidates one of the most common arguments for the permissibility of targeting enemy combatants, namely that they are in the business of harming the combatant doing the targeting.

What Zehfuss insufficiently credits here, however, is the collective conception of ‘risk’ and ‘threat’ that underpins the conventionalist account. Recall that war is a violent contestation between not only individuals, but also collective belligerents. Combatants are, it is presumed, simultaneously an individual material threat and a constitutive element of the collective threat of their side. This presumption of collectivised threat morally justifies the targeting of combatants on the basis of status, rather than the narrower and less permissive metric of conduct. The challenge to the conventionalist justification for lethal force is not, as Zehfuss suggests, inherent to the UAV. Rather, it is contingent on a level of military imbalance currently evident only in certain cases of UAV-exclusive violence. Within these specific conditions, the collective threat of the weaker side is mitigated to a degree that imperils the moral justification for status-based targeting.

The obvious should first be acknowledged: although UAV-exclusive violence has succeeded in insulating the military forces of the US from reciprocal harm, the American public remains under the general threat of retaliation in the form of terrorist attacks. ⁹ In this sense, it is perhaps more accurate to regard the physical risk sustained by the US as diffused, rather than transcended. ¹⁰ This aligns with Walzer’s (2004: 51–52) description of terrorism as the production of ‘general vulnerability’.

The extent of this enduring collective reciprocity, however, must be properly recognised. According to terrorism data sets, the likelihood of an American citizen being killed in a terrorist act is one in 3.5 million per year, in contrast to a one in 8000 chance of being the victim of homicide (Mueller and Stewart, 2012: 95–96). More recent studies rank the odds of death in a terrorist attack slightly above being killed in a tornado, and significantly less than being killed by the police (Mosher and Gould, 2017). ¹¹ Even if we factor in the generalised threat of terrorism, radically asymmetric UAV violence has mitigated the collective threat of those targeted to a degree far beyond prior warfare. Noam Lubell (2010: 117) has argued that the full tally of terrorist activity directed against the US over the past two decades, including the attacks of 11 September 2001, still only amounts to ‘isolated and sporadic’ violence.

The argument here is that the totality of this threat is insufficient to morally justify the use of status-based targeting within conditions of UAV-exclusive violence. ¹² This is enough to problematise the ‘global War on Terror’ standard set by the Bush administration, which justified the lethal targeting of suspected terrorists wherever they were located on the basis of group membership (Daskal, 2013: 1176). This does not mean that violence can no longer be exercised within these conditions. Those targeted by the US outside the active battlefield may very well be liable to direct and deliberate violence. Crucially, though, this determination should be reached on the basis of the target’s specific conduct. It is this individualist metric that is employed by Just War revisionists.

Revisionism opposes collectivised determinations of liability and, by extension, status-based targeting. In more recent years, the policy of the US, at least in relation to UAV-exclusive violence, has shifted to reflect this individualist account of enemy liability. ¹³ In May 2013, then President Obama (2013) issued a ‘presidential policy guideline’ limiting the use of UAVs outside active war zones to targets that present a ‘continuing, imminent threat to the American people’. This followed on from a 2012 statement in which Obama claimed that such strikes were restricted to threats that were ‘serious and not speculative’ (cited in The Economist, 2012). Former Secretary of State John Kerry echoed this, suggesting that UAV strikes were restricted to ‘confirmed terrorist targets at the highest level’ (cited in BBC News, 2013: emphasis added).

Even within the context of radical asymmetry, the killing of high-level enemies may be morally justified through an individualist, conduct-based metric of liability. Such individuals, it is reasonable to assume, play an essential role in the planning, operationalisation and execution of terrorist attacks. They are causally connected to a sufficient degree to the ‘continuing and imminent’ threat faced by Americans, even if geographically and temporally dislocated from the threat itself (Montague, 2012: 286–287; Tesón, 2012: 411–413). The question this immediately raises, however, is at what point does this assumption of liability become unreasonable? How are we to reconcile conditions of radical asymmetry with the killing of those whose contribution to the threat in question is more tenuous?

A 2010 Reuters investigation reported that of the 500 individuals killed by UAV strikes in Pakistan since 2008, only 14 were considered ‘top tier militant targets’, with 25 deemed ‘mid-to-high level’ (Entous, 2010). Low-level al-Qaeda and Taliban fighters were targeted and killed at a rate 12 times greater than mid- to high-level fighters (Entous, 2010). An evaluation of the Pakistan UAV program conducted the following year by the Obama administration found that low-level militants remained the disproportionate focus of strikes (Bergen and Rowland, 2012).

In Yemen, the figures are similarly troubling. By 2014, Central Intelligence Agency (CIA) and Joint Special Operations Command (JSOC) strikes, mostly via UAV, had killed between 500 and 1400 Yemeni targets (TBIJ, 2016). Although initially restricted to al-Qaeda’s senior leadership, targeting subsequently expanded to include local opposition to the authoritarian president, Ali Abdullah Saleh (Woods, 2015: 205). The US, argued Ryan Devereaux (2015), was ‘devoting tremendous resources to kill off a never-ending stream of nobodies’. The Obama administration maintained that it was conducting a ‘Just War’ in both Pakistan and Yemen. The nature of these strikes, however, suggests an over-permissive targeting criterion.

A crucial point to consider here is that these strikes would likely not be morally problematic — regardless of the individual threat level of the targeted fighter — if conducted in a more traditional, and higher-risk, war zone. Within such conditions, liability to harm may be justifiably established on the basis of general presumptions. Jennifer Daskal (2013: 1198) writes that in an actively contested battlefield:

[W]here troops on the ground are exposed to high levels of risk … those engaged in on-the-ground combat should not be required to hold their fire until they conduct a careful evaluation of the threat posed; such a rule would be potentially suicidal.

The permissiveness with which combatants may violently target one another on the battlefield is grounded in an assumption of reciprocal collective risk. Radically asymmetric violence undermines this assumption to a degree that calls the coherence of this permissiveness into question. Within such conditions, the presumption that low-level enemies are sufficiently liable for direct and deliberate targeting is no longer morally justifiable.

Issues of moral coherence are unlikely to significantly trouble those whose primary concern remains the strict legality of these strikes. As Gabriella Blum (2010: 72) observes, within mainstream interpretations of the laws of war, there exists a ‘general acceptance … of the near-absolute license to kill all combatants’. For as long as these laws remain applicable, this licence endures, no matter the degree of military imbalance. Before relegating the challenge of radical asymmetry to a marginal and exclusively moral concern, however, the critical and necessary interconnectedness between the moral and legal frameworks of war should be recognised. Moral judgement retains an essential role in the laws of war, functioning as an ongoing threshold test for their acceptability. According to Charles Garraway (2014: 88), just as morality without law is an aspiration, ‘the law of war without ethics is an empty shell’. To ignore this in the case of radical asymmetry is to risk a moral hollowing out of the laws of war. Radical asymmetry generates a problematic ambiguity, not only within the moral right to kill itself, but also in the relationship between this right and the legal guidelines of war.

It is also worth considering the significance of who exactly is doing the killing. Robert Jackson (2000: 173) defined ‘great powers’ as the ‘very select group of states whose policies and actions can affect the course of international affairs’. Although some view the trajectory of the US as one of relative decline, it remains the foremost of today’s great powers, with a global status that affords it a freedom of action that few other states can replicate. This status makes legitimacy challenges generated by its preferred mode of warfare particularly relevant.

Experts have long questioned whether American conduct during the War on Terror would trigger a ‘reverse cascade’ — a challenge to an existing norm that resonates with a relevant audience, causing the norm itself to lose salience (McKeown, 2009: 6). Attempts by the US to contest and reinterpret long-established rules concerning prisoner of war status and torture are but two examples of this process. More recently, Director of National Security and Human Rights at Amnesty International, Naureen Shah, criticised the Obama administration’s normalisation of the ‘bizarre’, in reference to what she perceived as the problematic scale and opacity of its UAV program (cited in Devereaux and Emmons, 2016). This line of inquiry must be broadened and deepened to include the tension between the increasingly one-directional nature of American violence and the foundations of Just War theory. Through its actions and policy, the US is currently in the process of establishing a normative and legal template that will inform the future use of radically asymmetric wartime practices. This article provides a theoretically grounded framework for better determining the point at which such violence is no longer compatible with the assumptions that underpin the moral right to kill.

Conclusion

As history demonstrates, during periods of profound organisational and technological change to the character of war, the degree and speed with which regulatory frameworks have adapted to reflect these changes has, at times, fallen short. An all-too-frequent consequence of this has been a reduction in the capacity of these same frameworks to properly evaluate and restrain military conduct. As early as 1920, political scientist James Garner (1920: 463) argued that ‘the existing [Hague] conventions are either silent, inadequate, or out of harmony with present-day conditions’. Similar explanations have been provided for the breakdown of mutual limits on violence during the Napoleonic Wars (Best, 1980: 76–77) and both world wars (Best, 1980: 76–77; Roberts, 1994: 124; Thomas, 2001: 89), as well as a number of post-1949 conflicts (Draper, 1998: 199). The moral equivalent of this disharmony can today be found in the context of radically asymmetric violence.

This article has examined three accounts of Just War: conventionalism, revisionism and contractarianism. The justification for inter-combatant violence advanced within each, it has been argued, is grounded to a significant degree in an assumption of war as a site of collective reciprocal risk between contesting belligerents. The erosion of this collective reciprocity, through conditions of radically asymmetric violence, represents a challenge to the very coherence of these accounts, particularly in terms of the permissiveness of inter-combatant violence. This article concluded with an overview of the moral challenge generated by the UAV-exclusive violence of the US. It argued that, within each of the three accounts of Just War, a tension can be located between the radical asymmetry of American violence and the assumptions that inform its permissiveness. This moral tension is most severe when violence is directed by the US against low-level enemies.

The challenge explored in this article is best understood as a ‘dilemma’, rather than ‘problem’. War is replete with problems — difficulties in need of resolution. Indeed, most war crimes fit this description. All too frequently, though, those who wage war are also required to navigate a range of moral dilemmas — situations in which no desirable outcome is obvious. What is under discussion here is not how best to prevent an objectively wrongful act committed by one party on the battlefield. Participants in war have long sought to neutralise both the threat-capacity of the other side and the risk exposure to their own. The drive towards risklessness in battle is logical, as dictated by the exigencies of war, and laudable, as dictated by the moral and political demands of force protection. This same drive, however, is moving us ever closer to a mode of violence that strains the basic foundations upon which our moral justifications for killing in war rest.