War and punitivity under anarchy

Abstract The individualisation of punishment is a key element in liberal narratives about international law and international relations. This narrative has become an integral part of positive international law, especially the regimes governing the use of force and the prosecution of an international crimes. Rather than punishing states or entire societies, liberals claim, punishment has become restricted to those who incurred individual guilt. To liberals, the individualisation of punishment is part of a larger process of enlightenment and civilisation that has helped to fence atavisms like revenge. We do not question the emergence of an ever more sophisticated system of individual punishment in international law. However, we argue that punitivity has been more difficult to fully channel towards individuals and away from collectives than claimed. To be sure, punitive language has by and large been banned from the laws of armed conflict. We argue, however, that the absence of a punitive vocabulary does not equal the absence of punitivity. In contrast, current state practices of using armed force are still imbued with punitivity, however silenced in the current legal framework and thus pushed underground. Realising the presence of a punitive undercurrent, we argue, adds to a more comprehensive understanding of contemporary state practices.


Introduction
On 7 April 2017 and on 14 April 2018, cruise missiles were fired into Syria by the United States and, in 2018, also by France and the United Kingdom. The 2017 strikes reportedly killed six soldiers and possibly also civilians. 1 This use of force was justified as a response to the use of chemical weapons a few days earlier, which, according to Western leaders, 'must not go unpunished'. 2 Similar language was used in April 2018: The White House referred to the use of chemical weapons as a 'crime' and added that 'with each chemical attack that goes unpunished, dangerous regimes see an opportunity to expand their arsenal', 3 whereas Theresa May stated that 1 At the time of writing, the number of casualties is clouded by the fog of war, with Syrian officials claiming nine civilians dead, including four children. See Harriet Alexander, Josie Ensor, and Roland Oliphant, 'US strikes on Syria: Xi Jinping told Donald Trump he understood the US response "because of the death of children"', The Telegraph (7 April 2017), available at: {www.telegraph.co.uk/news/2017/04/07/us-strikes-syria-tensions-rise-russia-warns-damage-ties-washington/} accessed 11 April 2017. The White House, 'Joined by Allies, President Trump Takes Action to End Syria's Chemical Weapons Attacks', available at: {https://www.whitehouse.gov/articles/joined-allies-president-trump-takes-action-end-syrias-chemical-weapons-attacks/} accessed 17 April 2018. the strikes were 'a clear message to anyone who believes they can use chemical weapons with impunity'. 4 The invocation of the vocabulary of punishment, retribution, and pedagogy deviates from the dominant liberal narrative about international law and international relations, which emphasises the individualisation of punishment. Rather than punishing states or entire societies, liberals claim, punishment has become restricted to those who incurred individual guilt. To liberals, the individualisation of punishment is part of a larger process of enlightenment and civilisation that has helped to fence atavisms like revenge. And indeed, international law by now has outlawed punitive wars, leaving only room for the use of force in self-defence, to enforce binding Security Council Resolutions and, controversially, to protect basic human rights. Punishment is delegated to the criminal justice system, which focuses on individuals, not states as such.
And yet, the language used in justifying military strikes against Syria is hardly surprising, as it fits in a longer pattern of post-1945 wars that have been justified in terms of 'punishment', or the need to educate deviant leaders ('teaching a lesson'). Apparently, punishmentthat is, the 'infliction of harm in response to a violation of a norm or rule' 5has been more difficult to fully channel towards individuals and away from collectives than claimed. The formal ban on punitive wars has not been followed by the absence of punitive justifications for armed interventions.
At first sight, this may lead to an obvious conclusion: since 1945 there have been regular violations of the prohibition of the use of force. From a formal, doctrinal perspective, this is indeed the case; the bombings of Syria in 2017 and 2018, for example, would hardly fit any of the doctrinally accepted exceptions set out in the UN Charter or customary law. However, as we will argue in this article, there is more to punitive wars than the question whether they stay within the strictures of the UN Charter. Paradoxically, it is the UN Charter itself, together with post-1945 developments in international law, which contains arguments that states have mobilised to justify punitive wars. This is not to say that these justifications are necessarily legally valid or morally convincing. Our aim is not to provide a substantive justification or assessment of the ways in which states legitimise their uses of force. Our aim is more modest: realising the presence of a punitive undercurrent, we argue, adds to a more comprehensive understanding of contemporary state practices and the different roles of law therein. The importance of adding this dimension can be illustrated by the recent discussion following the publication of Oona A. Hathaway and Scott Shapiro's The Internationalists. 6 As may be recalled, Hathaway and Shapiro argue that the 1928 Paris Peace Pact signifies a turning point in international politics, because 'for the first time in the history of the world, war was declared illegal'. 7 The Paris Pact was followed by subsequent attempts to limit recourse to war in the League of Nations Covenant and more radically, in the cornerstone of the UN Charter, the prohibition on the use of force in international relations. 8 Realists have taken issue with the claim that the legal prohibition on war or the use of force has actually contributed to a more peaceful world. 9 Students of the politics of international law have pointed out that the ban on war has a permissive dimension that defines 'other wars as lawfulspecifically, wars of "self-defense"'. 10  forms of doing politics conceptually possible. For example, with the establishment of the Security Council, new legal forms of violence were created as well. We build on this scholarship and focus on one particular driving force: punishment.
Our definition of our key concept, punishment, follows usage in sociology, 11 psychology, 12 and anthropology: 13 we define punishment as the 'infliction of harm in response to a violation of a norm or rule'. 14 This definition is deliberately broader than others by not incorporating two properties, namely: (a) that the norm in question has the status of a law; and (b) that the punishment is carried out by an authorised institution. 15 Both properties make sense in the context of a highly legalised setting of modern nation-states where punishment has by and large become 'third party punishment' 16 by delegating it to the criminal law system. For a study of 'punitivity under anarchy', however, excluding forms of 'second party punishment', that is, harm, inflicted by directly affected parties, risks missing one of the most interesting aspects, namely the degree to which punishment has (not) been delegated to third parties. As a consequence of including second party punishment, we do not distinguish sharply between punishment and revenge, which is typically carried out directly by the victim, but take the latter as a form of the former. 17 It is also worth noting that our definition is agnostic to the purpose of punishment and to the degree to which this purpose can be, or actually is, achieved. Whereas some definitions make retribution a defining element of punishment, we leave it deliberately open whether it is conceptualised in retributive, consequentialist, or expressive terms; the extent to which these rationales blend in punitive practices is best considered a question for future empirical research. 18 The consequentialist or utilitarian school of thought has much in common with theories of deterrence in International Relations: penalties are considered useful to deter unwanted behaviour in the future, with a view to both an individual person or state (individual deterrence) and to the community more broadly (general deterrence). 19 However, whereas punishment presupposes social norms whose violation it aims to prevent, deterrence is linked to the defence of national security interests, whose definition is by and large at the discretion of the deterring state and often does not include social norms.
In order to substantiate our main argument, we will proceed as follows. In the first section, we will set out why punitivity is likely to persist despite any formal prohibition of punitive wars. We will discuss research from various disciplines indicating that punitivity is at the same time hardwired and malleable. In the second and third sections we will zoom in on the two dominant forms of using and justifying armed force that populate the permissive space that the ban on war opened: wars of self-defence and humanitarian interventions. We use the 1986 bombing of Tripoli and recent threats to use force against Syria as empirical examples to illustrate our point. In both cases, we argue that sensitivity to a punitive undercurrent adds to a comprehensive understanding. We show that the political rhetoric in both cases was imbued with punitive language. In addition, we show that critics of the use of force made use of penal arguments, that is, they opposed the use of force because they doubted the underlying penal philosophy. Taken together, we demonstrate that an understanding of the terms of the debates would remain incomplete without due reference to punitivity. The concluding section addresses some normative concerns and consequences.
International law, punitivity, and the human condition There are good reasons to believe that punitivity is a hard-wired part of the human condition. Behavioural economists and evolutionary biologists argue that punishment is key to understanding how cooperation in larger groups is possible. Ernst Fehr and Simon Gächter coined the term 'altruistic punishment' to capture that the act of punishment does not provide any material benefit for the punisher but benefits the group by inducing potential non-cooperative members to cooperate. 20 Neuroscientists have supported this argument by demonstrating that 'people derive satisfaction from punishing norm violations' 21 and social psychologists have shown that these 'desires to punish are often the product of intuitive, rather than reasoned processes'. 22 Even though punishment may not serve their material interests, it is thus done because it feels right. Sociologist Émile Durkheim argues that punishment has a community-building function. 23 To Durkheim, the main function of criminal law is not to prevent harm from society and to deter potential wrongdoers. Rather, an important function of criminal law is to make society aware of the norms and values it shares and to reinforce them. According to Durkheim, punishing wrongdoers means to express the censure for the act committed. According to Morris Hoffman, 'we didn't really become civilized until we became willing to punish each other for generalized wrongs to each other'. 24 Although punitivity seems to be part of the human condition, there are obvious differences across time and space. Cultural psychologists emphasise that societies differ enormously as to what are considered punishable acts and appropriate penalties. 25 Sociologists and historians have documented the enormous changes in punitive practices in the course of human history. 26 One particularly interesting transformation concerns the depersonalisation of punishment. 27 Instead of victims often literally taking justice into their own hands, societies establish institutions and delegate punishment to increasingly specialised authorities. For liberal reformers like Franz von Liszt, the delegation of punishment to an independent third party is the essence of a modern, rationalised system of punishment. 28 The establishment of a monopoly of force and a Weberian state is a welcome side effect.
The depersonalisation and institutionalisation of punishment is particularly relevant in the context of our argument because it considers the punitive authority of societal institutions to result from an act of delegation: individuals forgo their punitive instincts for the sake of a rationalised system of punishment. By implication, however, the delegation of punitive authority can be revoked whenever the state is considered to fail in inflicting harm on those whose actions a society censures. The best-known examples are lynchings and other forms of 'popular justice' and vigilantism that occur where the state is weak or where its punitive practices are considered unsatisfactory. 29 As a consequence, state-sponsored institutions have an incentive to align punishment with popular expectations, even though this may conflict with a professional ethos of rationalised and 'civilised' punishment. Different practices of capital punishment across US states can be understood as such an attempt of the state to maintain the legitimate monopoly of violence (and thus punishment).
Taken together, our brief review of various disciplines' findings on punishment reveals three major points: First, punitivity is an integral part of the human condition. Although punitivity may have been expressed in different ways and to different degrees in international law and politics, it never disappears entirely. Second, punitive practices differ enormously across time and space. Third, punishment is often delegated to institutions that replace personal revenge with a rationalised form of punishment. Although celebrated as a civilisational act, the delegation of punishment is not irrevocable and may come under considerable pressure whenever punitive practices do not meet a society's expectations.
The three points made above help with understanding the development and transformation of punitivity in international law. 30 Under the rules prevailing in international law today, it is prohibited to wage punitive wars. While arguably 'armed reprisals' are still allowed in exceptional circumstances under jus in bello, the point of such reprisals is to stop ongoing violations of humanitarian law, not to punish the wrongdoer. 31 Under contemporary jus ad bellum punitive wars are simply prohibited. International law prohibits the use of force, with only two generally accepted exceptions: self-defence and authorisation by the Security Council, and one controversial possible exception, humanitarian intervention. This prohibition on punitive wars seems to be the logical outcome of a longer process in which the punishment theory of war became increasingly discredited. One of the obvious reasons for this process, already pointed out by classical just war thinkers, is the ever-looming risk of abuse and bias in the application of punitive force. 32 In addition, the idea of punishment between states was seen as irreconcilable with the principle of sovereign equality of states; a principle that since the seventeenth century grew into one of the foundations of modern international law. Yet, the prohibition of punitive wars does not do away with the deeply engrained desire to see transgressors of fundamental norms punished. Increasingly, international law sought to channel such punitive anxieties through criminal law, focusing on the criminal responsibility of specific individuals. Already after the First World War, the allies' request to have the German Kaiser extradited hinted at individual criminal responsibility as a new home for the punitive impulses. To be sure, the (in)famous war guilt clause in the Versailles Treaty and the reparations imposed on Germany demonstrate that collective retribution remained a strong motive on the side of the allied powers. Nevertheless, the focus on the German Kaiser already foreshadowed developments that would gain further momentum after the Second World War. After the Second World War, individual criminal justice was further institutionalised as an alternative to the punishment of entire societies. This development comprised two related elements. The first was that questions of (legal) guilt should be individualised, as expressed in the often-quoted adagio of the Nuremberg judgement, 'Crimes against international law are committed by men, not by abstract entities, and only by punishing individuals who commit such crimes can the provisions of international law be enforced. ' The second is that punishment should be about doing justice, not about pure revenge. This rationale was expressed, inter alia, by Nuremberg Prosecutor Robert Jackson who stated that the 'satisfaction of instincts of revenge and retribution for the sake of retribution are obviously the least sound basis for prosecution'. 33 Hannah Arendt even viewed retribution as 'barbaric '. 34 This idea formed the basis for the expansion of a system of universal jurisdiction, which granted states the power to exercise jurisdiction 'solely based on the nature of the crime'. 35 Both under treaty law and under customary international law, the number of crimes that could be prosecuted under the rubric of universal jurisdiction grew steadily, as the examples of 'genocide', 'torture', or 'crimes against humanity' indicate. In addition, it formed the basis for the establishment of international courts and tribunals, especially after the end of the Cold War. The 1990s witnessed a proliferation of special courts, ad hoc tribunals, mixed tribunals, and even a permanent international criminal court. Although the institutional settings of these courts and tribunals vary greatly, they all share a common assumption: guilt is to be assigned on an individual basis, according to legal proceedings. Blaming entire communities as well as seeking revenge should be prevented. 36 And yet, the question remains whether the rise of international criminal law has been sufficient to channel punitive impulses. Most violations of basic international law norms remain outside the scope of the criminal law system, both for jurisdictional and for practical reasons. It is even conceivable (although difficult to prove) that the rise of international criminal law has contributed more actively to the desire for some form of punitive response to violations of core values. Since the late 1990s the rise of international courts and tribunals go hand in hand with the propagation of the idea that there should be an 'end to impunity' for serious violations of international law. 37 Given that only a handful of cases are actually taken up in court proceedings, the ever-louder call to end impunity remains largely unanswered. This could help to explain why recourse to punitive language sometimes resonates so well in the justifications of wars. Typically, such wars are presented as wars against specific leaders or regimes (Saddam, Assad, Gaddafi, Milosevic) who have violated human rights and/or the prohibition of aggression. This fits in the A good example of this rationale can be found at the website of the International Criminal Tribunal for the Former Yugoslavia, which lists under the tribunal's main accomplishments that is has now shown that those suspected of bearing the greatest responsibility for atrocities committed can be called to account, as well as that guilt should be individualized, protecting entire communities from being labelled as 'collectively responsible'. The Tribunal has laid the foundations for what is now the accepted norm for conflict resolution and post-conflict development across the globe, specifically that leaders suspected of mass crimes will face justice. typical 'liberal way' of doing war: for the international community, for the population, and in order to eradicate evil regimes.
Be this as it may, it is clear that the weakness of a central authority through which punitive emotions can be channelled has put pressure on the strict prohibition of punitive war. Existing possible exceptions to the prohibition on the use of force have been invoked and stretched in order to justify punitive expeditions. Below we will give two examples, one pertaining to selfdefence, the other to humanitarian intervention.
Punitivity and self-defence: the 1986 Tripoli bombing and beyond In April 1986, the US bombarded several targets in Libya, killing an estimated 37 Libyans, including civilians. 38 Earlier that month, two Americans were killed and 78 injured when a bomb detonated in a Berlin nightclub that was known to be popular among American servicemen. The US government claimed to have conclusive evidence of Libya's responsibility of the attack. Libya's Colonel Gaddafi had indeed threatened terrorist attacks in retaliation for earlier confrontations in the Gulf of Sidra, which also led to the sinking of Libyan patrol boats. 39 The official legal justification for the 1986 American bombardment of Libya was self-defence against an ongoing pattern of attacks originating from Libyan direct and indirect involvement in international terrorism. The political justification was more ambiguous. On the one hand and in line with the individualised guilt discourse discussed above, the acts were individualised, as acts of Gaddafi and his regime, and separated from what the Libyan people must have wanted. On the other hand, however, despite the individualising rhetoric, the bombing itself did not hit Gaddafi personally but his troops and more than thirty civilians. Moreover, the Libyan acts against the United States were consistently labelled as criminal acts, not just as (imminent) armed attacks threatening US national security. The bombings on Tripoli, President Reagan stated, aimed at 'altering Gaddafi's criminal behavior'. 40 Last but not least, the United States presented the bombings on Tripoli as retaliation, a notion that followed from earlier discussions in the National Security Council, on 'retaliatory strikes' in response to terrorist attacks. 41 Debates in Congress also revolve around the pros and cons of punishment: Among others, Senator Denton argues that 'if prevention is not possible in a given case, the President must act to punish, while the gun is still smoking, the terrorist group or groups involved and the nationstates which have facilitated their criminal acts'. 42 More critically, Senator Hatfield observes that 'moral outrage' has 'risen dramatically as the United States watches terrorism go unpunished' and that this 'can be translated into punishment', adding, however, that 'today we are riding high but tomorrow, and next week, and next month, the consequences of our rash desire for retribution will have succeeded in rallying an otherwise factionalized nation behind its crazy leader'. 43 That members of parliament understand the punitive undercurrents of the strikes very well is also discernible in parliamentary debates in the United Kingdom. The arguments exchanged by supporters and opponents of punitive strike echo retributionist and rehabilitative arguments about domestic penal policy: 44 Supporters welcome 'some resort to punitive action' as 'the For the use of criminological language and notions in international security politics, see also Wolfgang Wagner, 'Rehabilitation or exclusion? A criminological perspective on policies towards "rogue states"', in Wolfgang Wagner, Wouter situation could not just be left to drift and slide' 45 whereas critics claim that the 'the strategy of using military force for the purpose of teaching Gaddafi a lesson is fundamentally flawed' because the 'use of such force will not prevent terrorism. Indeed, the use of such force is much more likely to provoke and expand terrorism.' 46 The Libya 1986 example is illustrative for the position of the right to punishment during the Cold War. Formally, there was no such right, with armed reprisals as a matter of jus ad bellum being effectively outlawed. Not surprisingly therefore, states justified their use of armed force primarily under the heading of self-defence, a notion that enjoyed worldwide acceptance, at least in the abstract. At the same time, however, the idea that some state leaders or regimes deserve to be taught a lesson was not abandoned. In the Libyan case, the attacks were justified not only as measures of self-defence, but also as forms of retaliation against a criminal regime that suppressed its own population and posed a threat of aggression to other nations. Retaliation was deemed necessary as a measure to prevent future criminal behaviour against the United States and its allies and, more generally, to change the behaviour of Gadhafi and his regime. This is not to say that the arguments of the United States were convincing to other states. The action was condemned by a majority in the UN General Assembly through Resolution 41/38, which spoke of an attack 'perpetrated against the cities of Tripoli and Benghazi'. 47 This shows that international law is not a one-way street: with its unilateral response to violations of basic community norms, the United States itself was branded as a state that trespassed peremptory norms of the international community. Still, the invocation of self-defence in combination with punitive arguments deserves further scrutiny. What is it about self-defence that made it such a natural ally for punitive actions? To be sure, the legitimisation of defensive wars in the UN Charter has a general permissive effect: 'to go ahead with a war, states are encouraged to explain how it qualifies as self-defense'. 48 In addition, it is useful to recall some reflections on the right to self-defence in early modern just war thinking. To begin with, self-defence was considered as a right primarily belonging to individuals, without the need for proper authorisation by a lawful authority (auctoritas). 49 The public element of 'authority' that is crucial for (classical) just wars is thus lacking. Furthermore, selfdefence had nothing to do with the core idea of just war thinking, the use of force to vindicate justice. Its foundation was the right to self-preservation, not some community ideal of justice. As a result, it was not necessary to show the existence of previous wrongful behaviour by another party. For self-defence to kick in, one needed to show the existence of a direct threat and the lack of alternative effective solutions, not the violation of some norm by the party against whom the self-defence operation was directed. Moreover, since everyone enjoyed the right to self-defence, it could also be exercised by persons who normally would be prohibited from participation in warfare, such as the clergy. Even those who participated in unjust wars retained their right to selfpreservation and thus to self-defence. Grotius, for example, argued that the right to self-defence could be exercised by 'every living creature', as it was grounded in the principle of self-preservation, not in See, for example, the treatment of self-defence in Francisco de Vitoria's 'De Indis', proposition three: 'Anyone, even a private person, can accept and wage a defensive war. This is shown by the fact that force may be repelled by force. Hence any one can make this kind of war, without authority from any one else, for the defence not only of his person, but also of his property and goods.' the vindication of rights or the need to punish aggressors and wrongdoers. 50 Most of the elements that make up classical just war thinking are thus absent in self-defence, with the notable exception of the well-defined object for which the war is fought (the so called Res). This picture changed fundamentally when self-defence was transposed to the collective level, to wars fought by public authorities. According to authors such as Gentili and Grotius, states enjoyed a right to defend themselves that went beyond the strictures of self-defence as a right enjoyed by everyone. Whereas individuals were only allowed to ward off actual or imminent attacks, states could also take broader anticipatory action, albeit within limits as to prevent that the defender himself would turn into an aggressor. The reason for differentiating self-defence operations from defensive wars conducted by public authorities, was that states do not live in a political society with courts and centralised enforcement. 51 It was therefore up to states to assume the role of judges and enforcers of the law. However, this also implied that it became much more difficult to distinguish defensive wars from offensive operations that sought to vindicate justice. And indeed, for Grotius, defensive wars often contained an element of law enforcement and even punishment. After all, defensive wars fought by public authorities not only dealt with threats, but equally with wrongdoings of the dangerous enemy. For that reason, defensive wars could easily turn into just wars in which 'it is permissible to forestall an act of violence which is not immediate, but which is seen to be threatening from a distance … by inflicting punishment for a wrong action commenced but not yet carried through'. 52 The right to self-defence in modern international law is very much like the war fought by public authorities, as discussed in early just war thinking. At first sight, it may look primarily as a right of self-preservation, meant to ward off an attack to one's state. 53 However, the set-up of the rest of the Charter and the development of international law has turned self-defence into something in-between self-preservation and the protection of community norms. The cornerstone of the UN Charter is Article 2 (4), which prohibits the use of force in interstate relations. Since its adoption in the UN Charter, Article 2 (4) has gradually obtained a status that goes beyond that of an ordinary treaty provision or a provision of customary law. As the International Court of Justice has argued, the prohibition on the use of force counts as a so-called peremptory norm of international law; a norm that reflects basic values of the international community and from which it is not possible to derogate. 54 In addition, the prohibition on the use of force is regarded as a so-called erga omnes norm. This means that all states and the international community as a whole have an interest in respect for this norm. This has several consequences for the understanding of self-defence.
For one, it lends support to a particular reading of the threshold of applicability for the right to self-defence. It has often been argued that states enjoy a right to self-defence only if the armed attack they suffer can be attributed to a state. 55 While this interpretation of self-defence  53 See, especially, the wording of Article 51 UN Charter, which gives states the right to self-defence in case of an armed attack. The article does not spell out from whom the attack should come and strictly speaking does not even require a previous wrongdoing on the part of another state before the right to self-defence can be exercised. is not uncontested, 56 many scholars and states still cling to the need to somehow attribute acts of private actors to a state. In response to the rise of trans-border attacks by non-state actors, the criteria for attribution have often been stretched and loosened, but few are willing to let go of the requirement of attribution altogether. 57 The rationale behind this argument is that self-defence operations necessarily imply the use of military force on the territory of another state. This would, under normal circumstances, be a violation of a peremptory norm of international law, the prohibition to use force in interstate relations. Only if states act in self-defence against attacks that can be attributed to another state, the argument goes, it may prove possible to circumvent the peremptory prohibition to use force: Using defensive force against the base of operations of NSAs [non-state actors] within a foreign host state's territory, even if that defensive force only targets the NSAs who have launched an attack, still amounts to a violation of the host state's territorial integrity. If Article 51 is to be a true exception to the prohibition on the use of force as set forth in Article 2(4) … it should respond in some way to the violation of the host state's territorial integrity. The legal mechanism which has traditionally been relied on to preserve an inter-state reading of Article 51, yet accommodate the need to respond to attacks by NSAs, is that of attribution. 58 However, if self-defence is turned into a response to previous wrongdoing, the line between law enforcement and defensive wars becomes thin, as already observed by Grotius. In this context, it is important to recall that only the more aggravated uses of force count as 'armed attacks' under international law. 59 States acting under self-defence are thus responding to more serious violations of a peremptory norm of international law. Often, such armed attacks will overlap with 'acts of aggression', which nowadays may even give rise to criminal responsibility for political leaders. 60 States acting in self-defence therefore do more than defending themselves or securing their subjective rights. Given the peremptory and erga omnes nature of the prohibition on the use of force, states acting in self-defence also provide a service to the international community as a whole. This picture would change if self-defence would be treated solely as a matter of necessity, based on the right to self-preservation. In that case, there would be no need for a previous wrongdoing; just an armed attack against a state would suffice. As the Tripoli example shows, however, in practice states or UN organs seldom invoke self-defence purely as a matter of necessity and survival. On the contrary: they generally also rely on 'pedagogical' arguments, the need to teach aggressors a lesson and the need to confirm to the world that the norm in international relations today is that of non-use of force.

Punitivity and humanitarian intervention: Responding to the use of chemical weapons in Syria
The community element is even more prominent when it comes to humanitarian interventions: interventions carried out to halt and prevent gross human rights violations, such as the ones in 56 For an alternative reading, see Thomas Franck, Recourse to Force, State Action Against Threats and Armed Attacks Kosovo in 1999. Humanitarian interventions are an almost perfect fit with the basic idea of just war theories: the offensive use of force for community purposes. Even in the nineteenth century, generally regarded as the heydays of positivism and state-centrism, humanitarian interventions remained important in international law and politics. As Stephen Neff has set out, they acquired the status of 'measures short of war', which required justifications that were directly borrowed from earlier just war thinking. 61 This all changed with the establishment of the United Nations. The UN Charter made an attempt to outlaw humanitarian interventions absent Security Council authorisation. It introduced a ban on the use of force in interstate relations, allowing for only two exceptions: the use of force mandated by the Security Council and individual or collective self-defence. The prohibition of aggression even grew into a peremptory norm of international law, as we have set out in the previous section. At the same time, however, international law codified an unprecedented number of humanitarian provisions, in the form of human rights treaties, provisions protecting civilians during times of armed conflict, provisions protecting future generations, etc. Several provisions in human rights and humanitarian law obtained jus cogens and erga omnes status, thus reflecting Immanuel Kant's adagio that 'a violation of rights in one part of the world is felt everywhere'. 62 Post-1945 international law thus pushed in two directions: (a) an unprecedented solid norm of non-intervention and non-use of force; and (b) an equally unprecedented codification of cosmopolitan norms that makes it everyone's business to care for basic human rights.
The increased humanitarian sensibilities help explain why humanitarian interventions kept popping up in international law, despite the prima facie clear prohibition contained in the UN Charter. This article is not the place to provide an in-depth discussion of the legal arguments that advocates and critics of humanitarian intervention have put forward. 63 Most international lawyers today would maintain that humanitarian intervention is (still) prohibited under rules of positive law, but also argue that in concrete cases there can be pressing moral concerns that would justify deviations from the strictures of the UN Charter. 64 As Bruno Simma has argued in relation to the threat of force against Yugoslavia in 1998: 'The October 1998 threat of air strikes against the FRY breached the UN Charter … But there are "hard cases" involving terrible dilemmas in which imperative political and moral considerations leave no choice but to act outside the law.' 65 In similar fashion, the Independent International Commission on Kosovo came to the conclusion that NATO's intervention was 'illegal but legitimate'. These arguments were not new to international law. Already before the Kosovo operation, authors had argued that (genuine) humanitarian interventions may be illegal, but should nevertheless be pardoned or 61 excused by states, or that humanitarian considerations should be regarded as 'mitigating circumstances' when assessing the violation of Article 2 (4) of the UN Charter. 66 These positions were partly derived from a reconstruction of the way in which states had responded to illegal humanitarian interventions, both during and after the Cold War. 67 The prohibition on the use of force thus has a somewhat paradoxical effect on the justifications for humanitarian intervention. Before 1945, it was possible to justify humanitarian interventions within the boundaries of international law. Post-1945, humanitarian interventions can only be justified if humanitarian concerns are so great and pressing that they justify bypassing international law. In other words: the need to rely on community values has only grown, precisely because of the codification of rules that seek to protect sovereign states against external aggression. Humanitarian intervention thereby becomes more than just saving lives and the rights of directly affected people. It is also an operation with high symbolic value, against an oppressor that violates the basic rights of humanity. Humanitarian interventions also serve as pedagogic tools, to set an example for all states where to draw the boundaries between normal conditions, when the use of force prevails, and abnormal conditions that justify recourse to formally illegal force.
The response to the 2013 use of chemical weapons in Syria is a case in point. The use of chemical weapons in a suburb of Damascus on 21 August 2013 was followed by threats of force by the US, the UK, France and other Western governments. Although they did then not directly lead to the use of force, they prepared the ground for the actual use of force for punitive purposes in 2017 and 2018. Thus, we think this episode reveals that humanitarian interventions may well have a punitive undercurrent. At first sight, the justifications for a possible attack on Syria in 2014 look very much like the justifications offered in the 1999 Kosovo intervention: while the action may be illegal under the strictures of the UN Charter, the use of force is legitimate as it seeks to protect core values underpinning modern international law. However, by contrast to the 1999 intervention, the bombings on Syria were openly defended as punitive actions. The 2013 debate thus brought together two types of arguments: one harking back to the right to punish as recognised in pre-1945 law and one invoking the 'illegal but legitimate' argument that was coined in the aftermath of the Kosovo intervention. The combination of the two strangely enough elevates the right to punish as somehow higher, or more innate, than positive international law.
Following the use of chemical weapons, calls for punishment were loudest in the media. On 31 August, The Economist's cover page read 'Hit him hard' and was followed by a lead article whose subtitle read 'present the proof, deliver an ultimatum, and punish Bashar Assad for his use of chemical weapons'. To be sure, governmental justifications to threaten or use force often did not rely on punishment only but more typically presented an argument that dressed punitive action in notions of (collective) self-defence or humanitarian interventions. In an interesting blend of selfdefence and punitive rationale, President Obama stated that This attack is an assault on human dignity. It also presents a serious danger to our national security. It risks making a mockery of the global prohibition on the use of chemical weapons. It endangers our friends and our partners along Syria's borders, including Israel, Jordan, Turkey, Lebanon and Iraq. 68 Although 'it also' (!) presents a danger to national security, the notions of 'assault', 'dignity', and 'mockery' rather point in the direction of a law enforcement officer who is expected not to tolerate blatant norm violations. Whereas the US President blended punitivity with self-defence, the British government presented a mix of humanitarian and punitive arguments. In the motion, introduced to the House of Commons on 29 August 2013, the government noted that the use of chemical weapons is a war crime under customary international law and a crime against humanity, and that the principle of humanitarian intervention provides a sound legal basis for taking action. 69 Along similar lines, the German Chancellor Merkel issued a statement that the Syrian government must not hope to continue violating international law unpunished. 70 That the 2013 Syria episode was more about punishment than humanitarian intervention is also supported by the manner in which the strikes were called off. Although legally not obliged, the British Prime Minister had decided to subject his policy to a vote in the House of Commons. To his surprise, a majority voted against the government motion, and he subsequently declared that 'the British Parliament, reflecting the views of the British people, does not want to see British military action. I get that, and the Government will act accordingly.' 71 After the British government backed off from military action, the French and American governments followed suit.
The debate preceding the vote had critically questioned the soundness of the humanitarian rationale for military strikes. MPs from both government and opposition parties doubted whether 'military action by our country would shorten the civil war', 72 'would cease the use of chemical weapons within Syria' 73 or would rather 'escalate violence either within the country or beyond Syria's borders?'. 74 Clive Efford wondered 'How do we ensure that no civilians die?' 75 Others criticised the punitive rationale directly: Conservative MP James Arbuthnot worried about a new doctrine of punishment as a reason for going to warnot deterrence, not self-defence, not protection, but punishment. I believe that, if that is a new doctrine, it needs considerably wider international consensus than currently exists. 76 Likewise, MPs wondered 'if we were to punish … every appalling regime by dropping missiles on it, would missiles not be criss-crossing the skies on a daily basis?' 77 Although the vote was certainly also informed by party politics, 78 the House of Commons debate can also be interpreted from a psychology of punishment perspective according to which the seven-69 and-a-half hours of debate cooled down the 'penal heat' 79 that had driven the government proposal for a military strike. Psychologists have pointed out that 'desires to punish are often the product of intuitive rather than reasoned processes' but that 'circumstances … may provoke the individual into reasoning about the case, and the reasoning system conclusion can override the response dictated by the intuitive system'. 80 The House of Commons debate did indeed critically examine the case for military action in Syria and exposed its weaknesses and inconsistencies. In the words of MP David Lammy, it had been 'impossible to have watched the footage in the past week and not to have felt the instincts of liberal interventionism pulsating in our consciences'. However, he warned of failure 'if we allow our thirst for justice to trump the patience to secure the greatest possible legitimacy for our action'. 81 Almost four years later, in April 2017, a new incident of chemical weapons' use was reported from the Syrian town of Khan Sheikhun. This time the US government reacted by firing 59 tomahawk missiles on the airfield from which the attack allegedly was conducted. In a letter to Congress, President Trump invoked 'vital national security and foreign policy interests' 82 as a justification, but the punitive overtones of the military action were difficult to miss: In a televised statement shortly after the attack, Trump stated that 'what happened in Syria is one of the truly egregious crimes. It shouldn't have happened. It shouldn't be allowed to happen … what happened in Syria is a disgrace to humanity. He's there, and I guess he's running things, so something should happen.' 83 The punitive language according to which the use of chemical weapons 'amounts to a war crime' 84 and that it 'cannot go unanswered' 85 was widely shared across Western political elites. 86 Whether such 'penal heat' is best channelled via international courts and institutions such as the ICC, the UN and the Organization for the Prohibition of Chemical Weapons (OPCW) or whether (Western) states should take justice into their own hands, was disputed among politicians in liberal democracies. Remarkably, however, there were few voices criticising the use of force for punitive purposes as illegal, illegitimate, and incompatible with the liberal project of individualising punishment, neither in 2017 nor in 2018.

Conclusion
Changes in the practices of states' use of military force have always attracted ample attention. Although the purposes for which states use armed force, the justifications they give and the ways they fight wars never converge completely, characteristic developments and patterns are often discernible. For example, practices after the Second World War have been described as a 'self-defence revolution' 87 because self-defence became a standard justification for using military force (while others, such as 'forcible 79 reprisals' went by and large out of business). Likewise, the post-Cold War era is often characterised as one of 'humanitarian interventions' and an emerging norm of 'Responsibility to Protect', both inspired by liberal conceptions of world order and resonating well with older notions of 'just wars '. 88 This article adds to this debate by setting out a constant factor in the use of military force by states. Throughout history, we claim, states have used military force not only to pursue their own interests or to save strangers, but also to punish other states that transgress what are perceived as basic norms of the international community. However, since 1945, it has become impossible to invoke punishment as such as legal justification to use military force in international relations. It is simply illegal under positive international law to assume the power to punish other states. And yet, punitive elements keep creeping into justifications for the use of force. What is more, punitive elements are constantly linked to generally accepted exceptions to the ban on the use of force, such as self-defence or to more contested justifications such as humanitarian intervention. As we have shown in this article, this is not only a matter of states cynically moulding the law to their interests. It is also a matter of international law offering room to link punitive justifications to more established exceptions to the ban on the use of force. The examples in this article focused on justifications offered by the United States and Britain. However, it would be a mistake to assume that the use of punitive justifications is solely an Anglo-American affair. After the 2015 Bataclan attacks, for example, president Hollande invoked the right to self-defence, in combination with the plight to be 'unforgiving with the Barbarians from Daesh', since France was 'foully, disgracefully and violently attacked'. 89 Although we do not go as far as Anthony Lang who argues that international society has become more punitive in recent years, 90 our findings challenge the liberal narrative according to which punishment has been successfully fenced and individualised. While we have no intention to diminish the achievements of the various international courts and tribunals, we question whether the 'penal heat' has been fully channelled from punishing collectives to punishing individuals. Instead, this article demonstrates that punitivity has been silenced in justifications of the use of force but not disappeared. A sensitivity to the punitive undercurrents of actual uses of force adds to a more complete understanding of contemporary practice.
If punitivity has indeed been more resilient than the liberal narrative claims, as we argue above, what consequences follow from this finding? Why does it matter? What are the normative implications? Rather than taking a strong position of our own, we conclude this article by mapping what we consider the main lines of argumentation that can inform this debate and by pointing out what future research should address. An obvious question for any normative assessment of punitivity in international politics is whether punishment works, that is, whether it makes norm violators stop violating norms and whether it has a general deterrence effect on third parties. While the effects of punishment in international relations merit further research, 91 a vast literature in sociology and criminology has accumulated evidence that at least in a domestic setting, punishment hardly ever makes those being punished norm-abiding citizens. 92 88 See also Anna Geis, Harald Müller, and Niklas Schörnig, The Militant Face of Democracy: Liberal Forces for Good (Cambridge: Cambridge University Press, 2013). 89 Instead, punishment has an effect on the community that carries out the punishment. As George Herbert Mead observed, the attitude of hostility toward the lawbreaker has the unique advantage of uniting all members of the community in the emotional solidarity of aggression. Furthermore, the attitude reveals common, universal values which underlie like a bedrock the divergent structures of individual ends that are mutually closed and hostile to each other. Seemingly without the criminal the cohesiveness of society would disappear and the universal goods of the community would crumble into mutually repellent individual particles. 93 Although Mead wrote with a view to domestic society, there is no reason to expect that the community-building effect of punishment should not be at work in international politics, too. This is another field that warrants further research. It would, for example, be interesting to examine to what extent community building depends on punishment being carried out by a second or a third party (such as the United Nations). Absent a central government, punishment in international politics is always vulnerable to critiques of bias and the insincere invocation of community norms. In such cases, punishment is more likely to divide than to unite the international community, as the examples of Tripoli (1986), Iraq (2003), or Libya (2011) have shown. Where those who intervene claim to protect basic community norms (for example, human rights), critics of the intervention invoke other community norms to brand the intervening states (for example, the prohibition of aggression).
If, however, we assume a community building effect and a limited, if any impact on norm violators, an assessment of punitivity will have to acknowledge that it is first and foremost about the community of the punisher(s). The main question then seems whether the emotional solidarity that Mead and Durkheim mention is worth the harm inflicted on a norm violator. An answer to this question will depend on the inclusiveness of the community (ranging from 'mankind', at one end of the spectrum, to the citizens of a particular country, at the other end), the moral dignity of the norm violated and the amount of harm to be inflicted to an individual or a collective.