Two Thoughts on Manuel Ventura’s Critique of Specific Direction

Manuel Ventura, the director of the Peace and Justice Initiative, has published two excellent posts at Spreading the Jam (here and here) that criticize the specific-direction requirement — and my defence of it. I cannot possibly address all of the points that Manuel makes, but I do want to respond to his understanding of the role that customary international law plays at the ICTY and his defence of the Special Tribunal for Lebanon’s (STL) analysis of the general definition of terrorism under customary international law.

Custom at the ICTY

As Manuel notes, the Special Court for Sierra Leone (SCSL) rejected Perisic‘s specific-direction requirement because it concluded that the requirement lacks an adequate foundation in customary international law. I criticized the SCSL’s position in a recent post, pointing out that the ICTY did not need to find a customary foundation for the specific-direction requirement:

Ad hoc tribunals are limited to applying customary international law because of the nullem crimen sine lege principle: relying on non-customary principles to convict a defendant would convict a defendant of acts that were not criminal at the time they were committed. The specific-direction requirement, however does not expand criminal liability beyond custom; it narrows it. There is thus no reason why the requirement has to have a customary foundation.

Manuel takes issue with my argument in an interesting way — by insisting that the ICTY can only apply legal principles that have a customary foundation, because customary international law is the only source of law that the Tribunal is empowered to apply:

But, says Kevin Jon Heller here and here, this is all irrelevant, as specific direction need not have a customary law basis since it only serves to narrow criminal responsibility rather than expand it. In his view only in the latter is nullum crimen engaged – the reason why the ICTY was mandated to apply customary international law. However, this view misses an important and very basic point. As he acknowledges, the mandate of the ICTY is to apply custom, and while it is true that nullum crimen is not engaged when criminal liability is contracted rather than expanded, it is also true that in not applying custom the ICTY is not applying the law it was specifically mandated and empowered by the UN Security Council to apply. If specific direction is not custom, then it is still applying something, but it cannot be called customary international law. In other words, it went beyond applying its governing law, and into a realm that is was not expressly empowered to go. In short, if specific direction is not customary, then it acted ultra vires and that is as problematic as a nullum crimen violation. It is not simply a bad policy decision that only engages ‘criminal law theory’.

There are two basic problems with Manuel’s argument. First, it is based on a misunderstanding of the ICTY’s mandate. Manuel claims that the Tribunal is empowered to apply one source of law and only one source of law: custom. But the Secretary-General’s report on SC Res. 808 does not say that. Here is the relevant paragraph about custom (para. 34):

The application of the principle nullum crimen sine lege requires that the Tribunal should apply rules of international humanitarian law which are beyond any doubt part of customary law so that the problem of adherence of some but not all states to specific conventions does not arise. This would appear to be particularly important in the context of an international tribunal prosecuting persons responsible for serious violations of international humanitarian law.

This paragraph does not say that every legal principle applied by the ICTY must have a customary foundation. It simply emphasizes that, given the gravity of international crimes and the importance of the nullum crimen principle, the Tribunal must scrupulously avoid convicting defendants of crimes that were not indisputably criminal at the time they were committed. Indeed, nothing in either the letter or the spirit of the paragraph suggests that the judges of the ICTY cannot narrow criminal responsibility when they believe the coherence of international criminal law or basic fairness requires it.

Second, and perhaps more importantly, the judges of the ICTY have never accepted Manuel’s interpretation of the Tribunal’s mandate. If the ICTY is empowered to apply only customary international law, the Tribunal’s entire jurisprudence concerning general principles of law is ultra vires, except with regard to the defences. Customary international law and general principles of law are autonomous sources of international law, and the Secretary-General’s report mentions general principles only in para. 58, which provides the Tribunal’s judges with the authority to draw upon general principles when deciding whether to recognize particular defences. Yet the judges have consistently relied on general principles to expand criminal responsibility beyond what they believed were the bounds of custom, holding — inter alia — that crimes against humanity demand a particularly severe sentence (Erdemovic), that non-sexual crimes do not have to be corroborated by witnesses (Tadic), that forcible oral sex qualifies as rape (Furundzija), that cumulative convictions are permissible (Kupreskic), and that individuals can be held in contempt of the Tribunal (Tadic). By Manuel’s logic, each and every one of those principles is invalid, because the ICTY is only empowered to apply customary law.

That cannot be correct. It thus seems clear that the best interpretation of the Secretary-General’s report is the one I defended above: namely, that custom is not the ICTY’s only source of law and is relevant only insofar as there is a question whether a particular legal principle is consistent with nullum crimen. As long as there is no potential nullum crimen violation, the Tribunal can apply both legal principles that expand criminal responsibility (via general principles of criminal law) and legal principles that narrow criminal responsibility (in order to maintain the coherence of international criminal law or because fairness demands it). The ICTY thus did not have to find a customary basis for the specific-direction requirement.

Custom at the STL

Manuel’s endorsement of the STL’s methodology for determining the content of customary international law is even more problematic. In its infamous Interlocutory Decision on the Applicable Law, the STL analysed whether a general definition of terrorism existed under customary international law. In conducting its analysis, the Tribunal acknowledged that there was a significant split in definitions of terrorism concerning the need for the the violent act in question to be committed for a political or ideological purpose. Some national jurisdictions and international conventions require such a purpose, while others do not. The Tribunal eventually concluded (1) that a general definition of terrorism does, in fact, exist under custom, and (2) that the general definition of terrorism requires only a violent act designed to spread terror or coerce a government, because although all national and international definitions require those two elements, they do not all require a political or ideological purpose.

In his post, Manuel endorses the STL’s analysis — and suggests it could help divine the customary definition of aiding and abetting:

I happen to think that the STL Appeal Chamber’s approach to customary formation – poorly understood by academia – could actually be of assistance. As I have argued here, the STL’s approach (in defining terrorism at international law) was to take various national definitions on terrorism and identify common elements among them – in substance, not form (replicating a similar approach as the (then) UK House of Lords in R v. Jones when it rejected the proposition that aggression was not sufficiently defined because at the time  – like terrorism today – its constitutive elements had not been comprehensively set out in a treaty). Having undertaken that review, the STL Appeals Chamber held that there was indeed a certain level of commonality, but this did not extend to an ideological or political motive element where state practice was divided (STL Applicable Law Decision, para. 106). Its solution? Hold that custom had emerged only with respect to the elements that were common among the practice of states. States with an ideological or political motive in their national definitions simply have a narrower understanding of the crime for the purposes of domestic law (just like many states define other international crimes differently at domestic law as compared to international law) or they could be seen as persistent objectors.

The STL’s approach to custom is not “poorly understood by academia.” On the contrary, scholars understand it all too well — which is why they have almost uniformly rejected it. Put simply, the decision is profoundly methodologically confused, as Manuel’s own summary of its analysis confirms.

The problem is the “common elements” approach itself. As noted, the STL was faced with two different specific definitions of terrorism in national practice and international conventions. Definition A has two essential elements:

1. Actus reus: a violent act.

2. Mens rea: intended to spread fear or coerce a government.

Definition B has three essential elements:

1. Actus reus: a violent act.

2. Mens rea: intended to spread fear or coerce a government.

3. Mens rea: for a political or ideological purpose.

The STL looked at these two specific definitions of terrorism and decided that the general definition of terrorism under customary international law includes Elements 1 and 2 but not Element 3, because only Elements 1 and 2 are shared by both definitions. There is, in other words, enough agreement between Definitions A and B concerning Elements 1 and 2 to generate a general customary definition, but not enough agreement to include Element 3 in that definition. The general customary definition of terrorism is thus Definition A, not Definition B.

But that is completely illogical. 

Why? Because it means that the STL counts states and conventions that adopt Definition B as evidence of a general definition of terrorism even though a significant number of acts that qualify as terrorism under the general definition would not qualify as terrorism under Definition B — namely, acts that are violent and intended to spread fear or coerce a government, but are not committed for a political or ideological reason. Such acts are simply not terrorism under Definition B. The existence of states and conventions that adopt Definition B thus cannot be evidence of a general customary rule that deems such acts to be terroristic.

The STL, in other words, got the analysis precisely backwards. Although not all acts that qualify as terrorism under Definition A qualify as terrorism under Definition B, the opposite is not true: any act that qualifies as terrorism under Definition B also qualifies as terrorism under Definition A, because Definition A does not consider terrorism to require a political or ideological purpose. The presence of such a purpose is simply irrelevant under Definition A. As a result, although there is not universal agreement that violent acts that lack a political or ideological purpose qualify as terrorism, there is universal agreement that violent acts that have a political or ideological purpose are terroristic. The only possible general definition of terrorism under customary international law, therefore, is one that requires a political or ideological purpose — Definition B. This “lowest common denominator” approach to custom is valid; the STL’s “common element” approach is not.

We can now understand why Manuel’s attempt to defend a customary definition of aiding and abetting that does not require specific direction is so problematic. Here is what he says:

Coming back to aiding and abetting, could this methodology not also be applied to various domestic definitions of aiding and abetting? In other words, if the practice of states are generally consistent (‘generally consistent’ comes from the ICJ in Nicaragua, Merits, para. 186, cited approvingly by the STL Appeals Chamber, where it rejected the notion that state practice must be absolutely rigorous) in substance with respect to aiding and abetting but for specific direction, one could argue that customary international law is actually consistent except when it comes to a specific direction element. All other things being equal, states that include some form of specific direction requirement (either in the actus reus or mens rea) simply have a different domestic understanding of aiding and abetting that goes one step beyond that under customary international law, or they are persistent objectors. In other words, state practice could actually support a customary definition that does not include specific direction.

This argument makes precisely the same mistake as the STL. Manuel is trying to reconcile two different specific definitions of aiding and abetting. Definition A has two elements:

1. Actus reus: an act that facilitates the principal perpetrator’s crime.

2. Mens rea: knowledge that the act will facilitate the principal perpetrator’s crime.

Definition B has three elements:

1. Actus reus: an act that facilitates the principal perpetrator’s crime.

2. Mens rea: knowledge that the act will facilitate the principal perpetrator’s crime.

3. Actus reus: specific direction of the facilitative act toward the principal perpetrator’s crime.

Manuel’s “common element” methodology says that the general definition of aiding and abetting under customary international law includes Elements 1 and 2 but not Element 3, because only Elements 1 and 2 are shared by both definitions. But the fact that a number of states adopt Definition B (which in the real world they do in terms of mens rea, by requiring the aider and abetter to intend the principal perpetrator commit the criminal act) cannot be considered evidence of a general definition that mirrors Definition A, because any act that lacks specific direction does not qualify as aiding and abetting under Definition B. In Definition B states, such an act is simply not criminal. By contrast, any act that satisfies Definition B also satisfies Definition A, because Definition A does not care whether or not the facilitative act is specifically directed toward the principal perpetrator’s criminal act. As a result, the general definition of aiding and abetting under customary international law — the definition over which there is universal agreement among states — must be Definition B, not Definition A. Definition B is the lowest common denominator.