Essence of Crimes against Humanity Raised by Challenges at ICC

Written by

Two recent challenges to the jurisdiction of the International Criminal Court in the Kenya situation bring out exciting questions of the essence of crimes against humanity. Defence counsel have challenged jurisdiction on the grounds that the violence in Kenya, which involved over one thousand killings and hundreds of rapes, did not constitute a crime against humanity (see: challenge, challenge and prosecution response).  This issue has divided the Pre-Trial Chamber in past proceedings.  As this is a jurisdictional challenge, the outcome can and almost certainly will be appealed to the Appeals Chamber.  The case raises extraordinarily difficult questions about the demarcation line between crimes against humanity and ‘ordinary’ crimes, and thus the role and scope of international criminal law.

Previously, in the decision to authorize the investigation, the Pre-Trial Chamber divided on this issue, with both the majority and the dissent providing compelling arguments. The ICC Statute requires a “State or organization” behind the crime against humanity; Judge Kaul in dissent argued for a more stringent standard of a “State-like” organizations, whereas Judges Trendafilova and Tarfusser adopted a more flexible “capacity” test for an organization.

The more stringent approach advanced by Judge Kaul has attracted support in thoughtful and well-reasoned recent scholarship.  My aim in this comment is simply to add that a convincing theoretical account can also be advanced on behalf of the majority’s broader approach.  I do not seek to point to any flaws in the reasoning of the dissent or the scholars favouring the more stringent approach, as their reasoning is perfectly sound and impeccable.  Indeed, I acknowledge that the “state-like” theory is internally coherent, reconcilable with limited doctrinal authorities and consistent with a sound theory of crimes against humanity.  My aim is modestly to lay alongside that theory another plausible theory, which is also internally coherent, consistent with the authorities and consistent with a sound theory of crimes against humanity.


The facts alleged in the Kenya situation involve a  political party enlisting an extensive criminal organization to attack rival supporters, with the passive complicity of the police.  Article 7(1) of the ICC Statute requires a “widespread or systematic” attack directed against a civilian population, and Article 7(2)(a) indicates that an attack requires a “State or organizational policy”.  In both the authorization of the investigation and again in the decisions to issue summons, the pre-trial judges differed over whether the “organization” requirement was met.  The dissent required a “state-like” organization, whereas the majority supported a more elastic “capacity” standard, featuring a list of factors.

The issue goes to the demarcation point between CAH and “common” crimes that do not fall within international jurisdiction. It requires us to clarify the nature and purpose of CAH, and whether and when organizations such as political parties, powerful terrorist organizations, or slavery rings can ever orchestrate a CAH.

There are at four major theories concerning the ‘policy element’: (1) that there is no policy element (advanced by Guénaël Mettraux and others and adopted in Tribunal jurisprudence), (2) that there must be a State policy (advanced by Cherif Bassiouni and others), (3) the theory requiring ‘state-like’ organizations and (4) broader theories encompassing organizations with ‘capacity’ to direct CAH.  I focus here on the latter two theories, because those are the positions reflected in the Pre-Trial Chamber decision and those are the two most plausibly available to the ICC.

Seen in this broader context, proponents of Theory 3 (state-like) and Theory 4 (capacity) agree on a great many issues.  We agree that CAH do not and should not include random crime waves, that a policy element is necessary to sustain the coherence of crimes against humanity, and that CAH may be initiated or endorsed by non-state organizations.  The remaining question is the standard required to constitute an organization”.

 Merits of the ‘state-like’ theory

The general approach by Judge Kaul in dissent has much to commend it.  I have  explored elsewhere a tendency, at least in earlier stages of ICL, for judges to take the “easier” route of adopting expansive and popular approaches without sufficiently rigorous analysis.  Judge Kaul’s reasoning was a credit to the Court, in staunchly resisting such temptation and adhering to a principled analysis.  I am in complete agreement with his principled methodology, including his scrupulous inquiry into drafters intent and sophisticated assessment of the purpose of the policy element.  While I agree with and applaud his methodology, I differ only in that these considerations may not have compelled quite so stringent a standard.

Thoughtful scholarly literature, including excellent articles by my friends Claus Kress and William Schabas, advances some meticulously-reasoned and compelling arguments as to the justification for the policy element, and in support of a comparatively demanding requirement of a ‘state-like’ organization.  Significant considerations include the need: (1) to focus on crimes that affect peace, security and well-being; (2) to focus on crimes less likely to be addressed and prosecuted by states (whether because the state is behind the crimes or because they are committed by a powerful group that the state cannot stop); (3) to focus on harms by entities who have a ‘responsibility to protect’; and (4) to ensure an appropriate jurisdictional/resource allocation so that the ICC focuses on the worst cases.  I have stated these considerations with brevity, and thus have not done justice to the arguments, but I consider each of the arguments to be plausible and sustainable. 

While scholarly attention to date has often favoured the dissent approach, I would suggest that the majority approach also offers a plausible and sustainable account, which warrants greater reflection. To explain why I believe the more flexible majority approach is sound, I first provide an account of the purpose and theory of the policy element.

The purpose and theory of the policy element

It is widely agreed that common crime waves, in which unconnected individuals act on their own initiative, do not constitute crimes against humanity.  An aggregate of crimes by individuals acting on their own initiative do not amount to an “attack”.  The logically corollary of the widely-accepted proposition (A) that an ‘attack’ cannot be the random acts of individuals acting on their own initiative, is (B) that the acts must be directed, instigated, or at least actively or passively encouraged by some source other than individuals.  The two propositions are the same proposition: the first is worded negatively (what an attack it is not) and the second is worded positively (what an attack therefore is).  Thus, I suggest that the policy element is simply the proposition that everyone agrees upon, expressed not as a negative exclusion but as the logically corollary positive requirement.

Skeptics often argue that a policy element is not required in order to exclude random acts of individuals, because the requirements of “widespread or systematic” already perform this function.  However, it can be demonstrated that they do not. Consider for example, a situation such as post-apartheid South Africa, which faced rampant crime, including 15,000-25,000 murders per year.  Assume that there is no state policy to encourage or condone such crimes; it is simply a situation of elevated domestic crime.  Yet if we mechanistically apply the requirements of crimes against humanity, without a policy element, we reach an odd result.  Murders certainly satisfy the requisite base crimes.  Fifteen thousand murders easily qualifies as “widespread”.  The victims of the crimes are civilians.  A person committing a serious crime in that context is aware of the surrounding context. Thus, all serious crimes committed in a context of “widespread” crime would constitute a crime against humanity.

Even opponents of the policy element would answer that such a scenario is not a CAH, because those unconnected crimes cannot constitute an ‘attack’ on the civilian population.  To which the response is, “Can you articulate the specific requirement in your definition of ‘attack’ that excludes those unconnected acts?”.  The answer to that question is the first key to the riddle of crimes against humanity.  The solution adopted in the Rome Statute, and significant other jurisprudence, is that the policy element provides the thread of connection between otherwise disparate acts, so that they may be described collectively as an attack directed against a civilian population.

 The argument for a modest standard for ‘organization’

If the purpose of the policy element is simply to exclude crime waves of individuals acting on their own initiative, then we do not need to import or impose any further adjectival limitations on the nature of the organization.  As long as there is some organized entity directing, instigating or encouraging crimes, then we are no longer confronted with mere spontaneous ‘crime waves’ and unconnected acts of individual wrongdoing.

David Luban’s illuminating piece, a “Theory of Crimes Against Humanity” provides helpful insights that can support this approach.  He argues that crimes against “humanity” represent an assault on our “humanness” and, in particular, our character as social and political animals.  Our nature compels us to live socially and to create organizations.  Crimes against humanity arise when those organizations attack us; they are “politics gone cancerous”.  Thus, “all humanity has an interest in criminalizing atrocious violence and persecution committed by organized groups against civilian populations.”  Luban’s account is elegant, and also offers a valuable distinction that genocide focuses on the collective character of the victims (and hence may not formally require a policy element), whereas crimes against humanity focus on the collective character of the perpetrators (and thus require state or organizational policy).

Theory 3 arguably focuses on betrayals of the responsibility to protect, ie. violence directed by structures with state-like power or authority.  One could argue that this imports too much of a state-orientation from human rights law, whereas international criminal law has developed separately, encompassing conduct orchestrated by states or non-state actors, and that we must recognize the massive harms that non-state actors can inflict.  Theory 4 accordingly focuses more simply on collectivized violence directed against civilian populations.   Crime waves of unconnected individuals or numerous spontaneous groups are “normal” crime, not engaging international jurisdiction.  But where an organization collectively orchestrates or endorses harm against a civilian population, and the harm reaches the level of “widespread” or “systematic” violence, then it is no longer “normal” crime but something more sinister, coordinated and dangerous, and warrants international concern. Thus, we can embrace the considerations offered above (such as the need to focus on threats to peace, security and well-being, and to ensure appropriate jurisdictional/resource allocation between national and international levels) and still find that such crimes warrant international prosecution where national systems fail to act.

In addition to this support in the theory (or in the object and purpose) of the policy element, other doctrinal considerations may also be invoked in favour of Theory 4.  Theory 4 is supported by the “ordinary meaning” of the term “organization”, as it does not require us to import significant additional requirements into the plain meaning of the term.  As long as an entity has sufficient institutional hallmarks to be called an “organization”, then we have a collective dimension as opposed to individuals acting on their own.  It is also supported by contextual interpretation, in that the structure of Article 7 features the disjunctive but high-threshold requirements of “widespread” or “systematic”, coupled with the conjunctive but low-threshold requirements of “multiple” and “policy”.  To pack substantial additional content into the “policy” requirement would arguably undermine the schema of Article 7.  Finally, with respect to drafting history, the policy element was very controversial and was unpopular with many delegations.  Many delegations reluctantly agreed on the element in order to rule out crime waves of unconnected crimes, and accepted the element on the understanding that it was a modest threshold that could be inferred from the circumstances. It would arguably be unfortunate to impose a significantly higher threshold in the name of fidelity to drafters’ intent, if an intent to insist on state-like entities has not been shown.

A narrowing of the gap

The gap can between Theory 3 and Theory 4 can be further narrowed.  I believe that the more stringent criteria laid down by Judge Kaul, and persuasively developed by others such as Claus Kress, do indeed have a role to play in the law of crimes against humanity.  I would suggest that they are requirements under the “systematic” branch.  Indeed, I believe this is necessary to make sense of the “systematic” branch, because it cannot be the thoroughness of planning alone that engages international jurisdiction.  Thus, I believe that the “systematic” branch focuses not only on the organized nature of the activity but also on the character of the entity organizing it.  The result would be that widespread serious crimes orchestrated by any organization would constitute a CAH (satisfying both “widespread” and “policy”) and would properly warrant international jurisdiction if states fail to prosecute. However, where a minor organization commits crimes on a small scale (not reaching “widespread”), then the mere fact of being meticulously planned does not convert those crimes (heists, assassinations) into CAH; they are still  just common crime.  Conversely, where a state or state-like organization methodically turns on a civilian population, in betrayal of the responsibility to protect, then this can constitute a CAH, even if the crimes do not reach a scale of “widespread”.  The state or state-like organization provides the requisite “system”.


My aim in this comment is to add to the excellent literature on Theory 3, by showing that Theory 4 (the majority approach and my own preferred theory) is also tenable and has much to commend it.  Inductive reasoning from doctrinal precedents does not resolve the question, because precedents at this stage are so sparse.  Deductive reasoning from theoretical constructs also does not resolve the question, because there are multiple plausible theories.  Accordingly, we may turn to a different form of inductive reasoning, testing the outcomes of competing theories in different hypotheticals against our intuitions, in order to try to unearth a deeper theory.  I for one would leave open the possibility that terrorist attacks on the scale of September 11 can qualify as CAH, as well as the possibility that an organized mafia could reach a scale of terror that would qualify, and that a slavery ring could reach a scale of operations that would qualify.  Accordingly, I suggest it would be unfortunate for the Appeals Chamber ICC to lock in too rigid a standard at this stage, and rather should leave room to develop the law with further  experience and reflection.

Print Friendly, PDF & Email


No tags available

Leave a Comment

Comments for this post are closed


Mathias Holvoet says

September 28, 2011

Thanks Prof. Robinson for this illuminating piece!
I have a question for you.
Could the crimes that are now under the jurisdiction of the STL qualify as crimes against humanity? There were plans to include crimes against humanity in the Statute, but those were blocked at the UN Security Council.
If I follow your theory I think this seems plausible: "widespread serious crimes (bomb attacks targetting politicians and journalists) orchestrated by any organization (presumably Hezbollah) would constitute a CAH (satisfying both “widespread” and “policy”) and would properly warrant international jurisdiction if states fail to prosecute (Lebanon asked to establish a Tribunal because of the inability of the Lebanese judicial system).

Or maybe you consider these assasinations as not widespread or systematic enough.? Anyway, the case of the STL again illustrates the looseness of the concept of crimes against humanity

jpaust says

September 28, 2011

Please note that customary crimes against humanity did and do not need a state or organizational policy, that genocide is an example of a subset type of CAH (and can be committed by private perpetrators on their own), and that the ICC simply has limited jurisidiction and cannot address all CAH under customary international law. See

tstevenson says

September 29, 2011

A review of this jurisdictional issue and why it is so important to the Kenya cases has also been posted to the ICC Kenya website. See

Darryl Robinson says

October 3, 2011

Thank you for these excellent comments.

Mathias' question about Lebanon is a very helpful one and points to the need to test out our different theories on some hard cases. At this stage it is difficult for me to answer simply because I still find the facts murky, even after reading several UN reports, so I would want to first reach firmer conclusions about the facts. But what you surmise about my view seems to be correct. From what I see, such as the allegations of 14 linked attacks, 'organizational policy' would not be an obstacle (if Theory 4 is correct); the issue will be showing 'widespread' or 'systematic'.

Professor Paust asserts an important view that there is no policy requirement for CAH at all. This view is noted in my post as "Theory One". It is indeed one of the plausible views about the policy element and a widely shared one. Much can be said in favour of this account. However I do not believe it is quite so settled as Professor Paust seems to suggest; there seems on the contrary to be considerable controversy. While great arguments have been advanced for Theory One, I personally happen to be more persuaded by the doctrinal arguments advanced by Schabas, Kress, Ambos & Wirth, Bassiouni and others. Moreover, the "no policy" account has to articulate a solution to the teleological and theoretical problems noted in my post and by the aforementioned scholars. So, I certainly respect the view but would humbly add that it it not quite so settled. In my view, as noted in the post, we have these multiple plausible theories and ICL discourse will continue to assess and refine them.

I also found the link in the third comment very informative -- thank you.

Meriam Z says

December 8, 2011

If couple of people can use nukes and kill thousands how is that not a crime against humanity? This number game is insulting to victims. I think intent is what matters. How many people have to get together to commit a crime against humanity.
We need harsher sentences too. A 27 year prison for Momcilo Perisic was a slap in face. I wwas hoping for life behind bars.
Howwever, it would be interesting to see what would happen if Croatia and Serbia sue each other at ICJ (different than ICC but both international)

Darryl Robinson says

December 8, 2011

The point of my post was precisely that - a small group of non-state actors inflicting thousands of deaths may indeed be a crime against humanity.

Meriam AZ. says

December 9, 2011

Thanks! makes sense but How many people have to die before it becomes a crime against humanity? Thousands?

Darryl Robinson says

December 14, 2011

It is a good question. If one relies on the "widespread" branch, then there is indeed an objective requirement of scale. The threshold is not necessarily thousands, however. There was an ICTY case that found a "widespread" attack with around 350 victims. Under the "systematic" branch the focus is not on scale but rather on the method and organizing entity, and hence the numbers could be much lower.

This may all sound rather technical. Indeed, all crime and all victimization is of course an affront to victims and to the community. The contextual element of CAH is a rule that allocates jurisdiction: it delineates when that wrong can be judged at the international as well as national level.

Mathias Holvoet says

December 16, 2011

If the contextual element is meant to allocate jurisdiction, can it be argued then it is somehow related with the principle of complementarity?
If there is State policy to attack a civilian population, like nowadays in Syria, there will be quite logically not much willingness to prosecute any crimes committed by the State, so international jurisdiction is warranted in these cases. Furthermore, if the crimes comitted by an organization are so widespread or systematic so that it make a national judicial system unable to cope with, international jurisdiction seems also necessary. Lebanon is again a good example here. They were unable to investigate and prosecute the crimes themselves, so they requested UN-assistance. Maybe it would have been better if they had 'self-referred' the situation by an ad hoc acceptance of the jurisdiction of the ICC, who would then be able to prosecute the crimes as CAH, instead of establishing another costly UN-tribunal.

Nadia NAM says

December 26, 2011

When states kill their own people then it should be a CAH. It seems reasonable to prosecute heads of state like Assad for CAH.
Genocide can be committed by few private perpetrators on their own, as mentioned by Paust, but it seems that it’s really hard to prove this crime. Momcio Perisic was never convicted of genocide yet he clearly wanted to destroy Bosnians. Assad might never get convicted of this crime either.

Nadia NAM says

December 26, 2011

When states kill their own people then it should be a CAH. It seems reasonable to prosecute heads of state like Assad for CAH.
Genocide can be committed by few private perpetrators on their own, as mentioned by Paust, but it seems that it’s really hard to prove this crime. Momcio Perisic was never convicted of genocide yet he clearly had the intent to destroy Bosnians. Assad might never get convicted of this crime either.