Syria and the Semantics of Intervention, Aggression and Punishment

Written by

untitledCarsten Stahn is Professor of International Criminal Law and Global Justice and Programme Director, Grotius Centre of International Legal Studies, University of Leiden

As I have tried to argue in a draft paper to be published in  the Journal of International Criminal Justice (Vol. 11, Issue 5) one of the most striking features of discussion concerning the legality of strikes against Syria has been the mixture of semantics relating to intervention. The justification of intervention in Syria has been replete with references to notions of accountability, deterrence and punishment. For example, US Secretary of State John Kerry linked the moral legitimacy of the proposed response to the idea that ‘there must be accountability for those who would use the world’s most heinous weapons against the world’s most vulnerable people’.  Also, the United Kingdom Legal Memorandum speaks of the objective of ‘deterring or disrupting the further use of chemical weapons’. Section 2 of the Resolution of the US Senate Foreign Relations Committee invokes three rationales of the authorization of military intervention: (i) ‘respond to the use of weapons of mass destruction by the Syrian government’; (ii) ‘deter Syria’s use of such weapons’; and1 ‘degrade Syria’s capacity to use such weapons in the future’.

This language associates military intervention with notions of ‘retribution’ and  ‘punishment’. The justification seems to be predominantly linked to the ‘gravity’ of the violation. The Obama Administration has translated this imagery into the notion of the ‘red line’. The  metaphor seems to suggest that such a breach of international law is so severe that it cannot be tolerated by the international community and warrants a repressive response. The concept that certain acts might trigger international responsibility and are of concern to the ‘international community as a whole’ is of course well-known (ICJ,  Barcelona Traction) and reflected in the ILC Draft Articles of State Responsibility (Art. 48).

The use of chemical weapons against civilians is both a violation of customary law and an international crime (UK Legal Memorandum, para 2).  But the consequence of this approach, i.e., the idea to ‘punish’ a regime for unlawful action under jus in bello through resort to the use of force, is highly controversial. There are other competing, and at least similarly important ‘red lines’ that require respect, i.e., Articles 2 (1) and (4) of the UN Charter, the prohibition of aggression and the viability of the collective security system, as argued by others.

The policy case in favor of intervention mixes two levels of discussions that have been separated in past decades: state responsibility for an international wrongful act, and individual criminal responsibility for international crimes. The idea that another state (rather than individuals) might be ‘punished’ has been banished from contemporary international law, in contexts such as sanctions, the ban on reprisals against civilians or the prohibition of collective punishment in international criminal law (e.g., Article 4, paragraph 2(b), of Additional Protocol II to the Geneva Conventions), since it has an indiscriminate effect. The current discourse on intervention appears to re-open this principle.

Humanitarian Intervention v. Protection of civilians

The idea to hold a regime accountable for crimes through military intervention remains one of the most contested propositions under international law. Two of the main theories that have been invoked are ‘humanitarian intervention’ (see the posts by Daniel Bethlehem, Dapo Akande and  Guglielmo Verdirame)  and ‘protection of civilians’.

‘Humanitarian intervention’ has its roots in the ‘right intention’ requirement of just war theory. It is part of the vocabulary of justification under jus ad bellum since the second half of the 20th century. It has been used to provide both legal and moral authority for action, in light of the motives and consequences of intervention. In cases such as Afghanistan and Iraq, it has been mixed with other claims, such as self-defence,  collective security (i.e., Security Council endorsement or implicit authorization) and/or pro-democratic motivations in order to justify action against a ruling regime.

Protection of civilians (PoC) is a jus in bello concept by origin. In recent, years, it has gained ground in the context of the mandate of peace operations. In the context of Libya, it has been included in the authorization language for enforcement action under Chapter VII in para. 4 of SC Res. 1973 (2011)), which was interpreted by some powers as a title to take action against the Gaddafi regime. In the Syrian context, it is again invoked for similar purposes. Para. 3 of the UK Legal Memorandum contemplated a modified version of the Libyan formula for a Security Council authorization. It suggested ‘seeking a resolution … under Chapter VII’ which would ‘authorize member States, among other things, to take all necessary measures to protect civilians in Syria from the use of chemical weapons and prevent any future use of Syria’s stockpile of chemical weapons’.

Both concepts seem ill-equipped to accommodate goals of ‘punishment’ and regime accountability.  They are meant to increase moral authority for intervention. The concept of PoC has visible attraction, since it focuses on goals that may be more easily attainable than broader objective of ‘humanitarian intervention’ (i.e., sustainable peace, transformation beyond the status quo ante).  But both concepts have a victim-centred, rather than a perpetrator-based focus.  They are guided by the rationale of protection of victims of conflict. Regime accountability may be a consequence or effect of protection. But it is not at the heart of the rationale of justification.

Current discourse seems to turn this logic on its head. It uses protection as a means to achieve accountability through military force. This approach may effectively weaken the impetus and credibility of humanitarian law and further erode the virtues of the separation of jus ad bellum and jus in bello.

Reprisals, countermeasures and ‘crimes of states’

As argued by Shane Darcy, the claims for justification of strikes bear most resemblance with other instruments of law that have been restricted in the second half of the 20th century, namely reprisals or countermeasures.

Reprisals have been a traditional method of enforcement of international humanitarian law. An armed reprisal in peacetime (peacetime reprisal) is generally considered prohibited. ‘Belligerent’ reprisals have been gradually constrained through treaty law and international practice, in particular in relation to protected objects and persons under the Geneva Conventions and Additional Protocol I. In the Kupreskić case, the ICTY Trial Chamber famously held that reprisals are subject to international rejection since they are ‘inherently a barbarous means of seeking compliance with international law’ (para. 528). Rule 145 of the ICRC Customary Law Study states that ‘[w]here not prohibited by international law, belligerent reprisals are subject to stringent conditions’. Several states have made reservations to the 1925 Protocol for the Prohibition of the Use of Asphyxiating, Poisonous or Other Gases, and of Bacteriological Methods of Warfare, in which they reserved the right to retaliate in cases of a breach of the declared Protocol.

But there are several problems with the applicability of the doctrine of reprisals in relation to strikes against Syria. Neither the UK nor the US are parties to the conflict in Syria. An armed reprisal would thus a priori constitute a prohibited peacetime reprisal. Moreover, as argued by the ICRC Customary Law Study, reprisals ‘may not be carried out for the purpose of revenge or punishment’. Any forcible action to seek compliance with the law would be a new type of reprisal, based on a breach of an ‘obligation owed to the international community as a whole’.

This idea stands in contrast to existing international law. Forcible countermeasures are ruled out by Principle 1 of the Friendly Relations Declarations, and the Draft Articles on State Responsibility of the ILC. Art. 50 makes this explicit in several ways. It states that countermeasures ‘shall not affect’:

‘(a) the obligation to refrain from the threat or use of force as embodied in the Charter of the United Nations; … (c) obligations of a humanitarian character prohibiting reprisals; (d) other obligations under peremptory norms of general international law.’

Punitive strikes would violate these principles. They would come close to retaliation action.

The existing rhetoric appears to tie intervention to the old argument that states may commit ‘crimes’. This approach was discarded by the ILC in the Articles on State Responsibility. The International Court of Justice confirmed this in the Genocide case (Judgment, 27 February 2007). It held that ‘as a matter of general principle, international law does not recognize the criminal responsibility of the State’ (para, 170). It argued that the Genocide Convention ‘does not provide a vehicle for the imposition of such criminal responsibility’ since obligations under the Convention are ‘responsibilities under international law’ that ‘are not of a criminal nature’. If the Genocide Convention does not contemplate such a logic, it would be difficult to invent it in relation to violations of the prohibition concerning chemical weapons.


The semantics of intervention have gained yet another level of complexity with the ongoing discourse on the definition and boundaries of aggression. Kampala Resolution RC/Res. 6 has brought the prohibition of aggression back from history books. The definition distinguishes prohibited acts of state, i.e. the ‘act of aggression’, from the notion of the ‘crime of aggression’. It requires an act which ‘by its character, gravity and scale, constitutes a manifest violation of the Charter’ (Art. 8bis (1)).  This discussion has a spin-off effect on contemporary debates on the use of label of aggression in relation to intervention.

Syria poses the ‘hard’ question of when and under what circumstances intervention would fall under the label of aggression. At Kampala, the US delegation lobbied in favor of an Understanding that explicitly excluded ‘humanitarian intervention’. It stated:

‘It is understood that, for purposes of the Statute, an act cannot be considered to be a manifest violation of the United Nations Charter unless it would be objectively evident to any State conducting itself in the matter in accordance with normal practice and in good faith, and thus an act undertaken in connection with an effort to prevent the commission of any of the crimes contained in Articles 6, 7 or 8 of the Statute would not constitute an act of aggression.’

This Understanding failed to gain support. The mere rejection does not imply that ‘humanitarian intervention’ is intended to come within the ambit of aggression. However, it makes the assessment considerably more difficult. It requires substantiation.

It is clear from the Understandings that a violation the Charter as such is not sufficient to qualify conduct as aggression. Understanding 6 clarifies this point, noting that ‘aggression is the most serious and dangerous form of the illegal use of force’. It specifies that the determination requires a holistic assessment. It ‘requires consideration of all the circumstances of each particular case, including the gravity of the acts concerned and their consequences, in accordance with the Charter of the United Nations’.

The test is predominantly an objective one. The question as to whether a violation of the Charter is ‘manifest’ is a qualitative determination that needs to be assessed in light of three criteria which apply cumulatively according to Understanding 7, namely the ‘character’ of the act, its ‘gravity’ and its ‘scale’. The Introduction to the amendment to the Elements of Crimes clarifies that ‘manifest’ is an ‘objective qualification’. The text leaves it ambiguous to what extent these criteria relate to modalities of action or legal qualification, i.e., the scope of the violation. Understanding 6 seems to suggest that they relate to underlying patterns of action, i.e. the nature and degree of use of force.

Legally, there are various possibilities to argue that specific types of ‘humanitarian interventions’ do not qualify as aggression. Both their ‘character’ and their impact (‘consequences’) might be used to distinguish them from other types of intervention (see e.g.,  Beth van Schaack). Moreover, arguments relating to the degree of legality might come into play, especially if ‘gravity’ is understood as a normative criterion, i.e. as one relating to the scope of the breach of law, as argued by some.

But it is equally clear that mere motivation or ‘good intention’ is clearly not enough. The fact that a State claims to act in ‘good faith’, i.e. for humanitarian purposes or to remedy a breach of a fundamental norm, would not per se exclude it from the scope of application of the Kampala amendments.  Calls to define aggression as a specific intent crime, based on aggressive intent or purpose (i.e., with the aim of occupation, subjugation or annexation) were not successful (see inter alia Ambos). The language of criminal law, as ambiguous as it is, might thus ultimately strengthen the prohibition of the use of force.


The Syria debate folds criminal justifications into the rhetoric of intervention. Following crime-based logic, the most natural step would have been to apply the principles and response schemes of the ‘Responsibility to Protect’, including a referral of the situation to the International Criminal Court, as suggested by the Special Adviser of the Secretary-General on the Prevention of Genocide. But formal ‘justice’ approaches and institutions have been sidelined in discourse. Instead, the focus has shifted to intervention. This is a dangerous tendency. Use of force cannot and should not serve as a short-cut to international justice or as a means of punishment. Examples, such as Iraq, have shown that frameworks and decision-making processes relating to the use of force to do not provide sufficient room to examine and test evidence in relation to violations and accountability. Any extension of the concept of intervention to ‘sanction’ violations of fundamental norms through military action would mark a serious step back. It would mark a departure from foundational principles of the international legal order that have been established over decades. Such a precedent might polarize debate even further, if criminal labels are attached to unlawful uses of force. With emerging evidence on the modalities of the use of chemical weapons (e.g., by UN inspectors), instruments of international criminal justice will need to come back to the forefront of debate, as suggested by the Secretary-General.

Print Friendly, PDF & Email
  1. iii []


No tags available

Leave a Comment

Comments for this post are closed


Jordan says

September 19, 2013

Carsten: you are quite right that "reprisals" and "punishment" as sole justifications have no traction under interntional law regarding the use of armed force. However, the Obama Administration might claim other types of justification in the future and "mixed motives" would not obviate legality based on an approprite claim. Ihave suggested that there might be seven claims, although some would depend on future events. See prior responses here re: an essay in JURIST.
Also, you rightly note that a violation would have to be, at least, a "manifest" violation of the Charter for a given use of force to constitute a criminally-sanctionable "aggression." This point is relevant with respect to some of the possible legal justifications for use of force in Syria, since some might have their detractors but, even if they would not ultimately be acceptable, a given use of force based partly on a controversial claim might not constitute a "manifest" violation of the Charter.