All posts by Thomas Weigend

Mr. Bemba’s Acquittal – A Shortcut to Justice?

Thomas Weigend is a Professor of International, Comparative and German Criminal Law at the University of Cologne.

(1) Introduction

The source of Mr. Jean-Pierre Bemba Gombo’s legal problems was an intervention of troops under his command in an internal armed conflict in the Central African Republic (CAR) in the years 2002 and 2003.[1] While Mr. Bemba continued to reside in the Democratic Republic of Congo (DRC), his soldiers committed murders and rapes against CAR civilians and pillaged their homes. Several years later, the Prosecutor of the ICC charged Mr. Bemba with the crimes committed by the soldiers, under the heading of command responsibility (Art. 28 ICC Statute). After having been convicted by a Trial Chamber in 2016, Mr. Bemba appealed, alleging that the Trial Chamber had committed a long list of factual and legal errors. The Appeals Chamber, by a 3:2 majority, found in favor of the appellant and acquitted him.

It is unusual, even in the area of international criminal law, that the resolution of a case on appeal engenders four separate opinions with an aggregate of about 500 pages. This avalanche of words is the result of a deep split between the judges of the Appeals Chamber – a split that affects almost every aspect of the case and has led to widely diverging views as to its appropriate resolution.  

It is impossible in a short comment to do justice to the multitude of arguments exchanged between the opposing sides as well as to the profound general observations and learned quotations in the separate concurring opinion of Judge Eboe-Osuji, the current President of the International Criminal Court. I will instead limit my remarks to what I conceive to be the main points of dissension among the judges.

Judges Van den Wyngaert and Morrison, who together with Judge Eboe-Osuji formed the majority, wrote a separate opinion[2] in which they explained their motives. In concluding this opinion, they seek to style the main conflict as one between reason and emotion:  “We can only hope to establish the rule of law”, they write, “if we discipline ourselves to be guided by rationality and resist the urge to allow emotions to determine judicial decisions”.[3] Although this sentence is meant to explain that the judges in the majority empathise with the victims of the atrocities committed by Congolese soldiers in the CAR, it implicitly portraits the dissenting members of the Appeals Chamber (Judges Monageng and Hofmanski) as following their emotions rather than rationally applying the law. Such an implication would be far from accurate, given the meticulous analysis of the evidence and the law that the dissenting judges provide in their lengthy opinion. Yet, it is hard to deny that the end result of the case leads to a sense of frustration – not only in regard of the enormous amount of time and effort spent on prosecuting and adjudicating Mr. Bemba but also in regard of the great suffering that is left without a legal response, at least on the level of international criminal law. The outcome of the case is unsatisfactory in yet another respect: although the Court was eventually unable to establish facts indicating Mr. Bemba’s guilt beyond a reasonable doubt, some of the judges remain fully convinced that the former accused was responsible, as a commander, for the crimes committed by DRC soldiers in the CAR. This leaves the case (and the former defendant) in a state of limbo, so that the efforts of all participants to establish “the truth” seem to have been in vain.

(2) Breadth of Charges

In the end, this problematic result has been caused by evidentiary difficulties. But as a preliminary matter, the very scope of “the case” before the Court was doubtful. Both the Trial Chamber and the dissenting judges of the Appeals Chamber accepted the charges as formulated by the Prosecution. These charges were formulated in very broad terms. For example, the charge relating to the crime against humanity of murder read: “From on or about 26 October 2002 to 15 March 2003, Jean-Pierre Bemba committed crimes against humanity, by the killing of men, women and children civilians in the Central African Republic […].” The document containing the charges listed a number of examples of specific criminal acts but made it clear that the charges “include, but are not limited to” those acts.[4] The dissenting judges argue that the formulation of the charges lies in the discretion of the Prosecutor, hence all acts of murder committed on the territory of the CAR between 26 October 2002 and 15 March 2003 by any soldier under Mr. Bemba’s command was properly before the Court.[5] As a consequence, the Prosecution was free, in the opinion of these judges, to augment the charges by adding further instances of crimes allegedly committed by the DRC soldiers after the original charges had been confirmed by the Pretrial Chamber and even after the trial had begun, provided that the defendant was given proper notice in accordance with Art. 67 (1) (a) ICC Statute. The Prosecutor had in fact amended the original charges in the course of the proceedings. The majority of the Appeals Chamber find that these additional instances – for which the Trial Chamber had held Mr. Bemba responsible under Art. 28 ICC Statute – were not subject to the Trial Chamber’s jurisdiction. The majority rely on the wording of Art. 74 (2), second sent. of the Statute, which reads:

“The decision [of the Trial Chamber] shall not exceed the facts and circumstances described in the charges and any amendments to the charges.”

The interpretation of this provision may be aided by Regulation 52 (1) (b) of the Regulations of the Court:

“The document containing the charges referred to in article 61 shall include: (…) (b)  A statement of the facts, including the time and place of the alleged crimes, which provides a sufficient legal and factual basis to bring the person or persons to trial, including relevant facts for the exercise of jurisdiction by the Court; (…).”

From these legal texts, the Majority of the judges draw the conclusion that the charges to be confirmed by the Pretrial Chamber must be precise in describing the acts of which the defendant is suspected; it is not sufficient that the document containing the charges simply lists categories of crimes (such as “murder” and “rape”) or – as happened in Mr. Bemba’s case – names a time span of almost five months and a space covering a whole country, thus holding the defendant potentially responsible for any offense committed by the troops under his command within these broad limits.[6] The Majority therefore accept as properly charged only those few particular instances of soldiers’ criminal conduct that were specifically named in the original charging instrument (one murder, the rape of 20 persons and five acts of pillaging).[7] In their Separate Opinion, Judges Van den Wyngaert and Morrison clarify the reasons for this conclusion: “The proper identification of criminal acts in an exhaustive manner is crucial to allow a Trial Chamber to manage the trial proceedings and to allow the accused to prepare a meaningful defence…”[8]

In my opinion, the last sentence quoted above correctly defines the issue. Contrary to the Minority’s argument, Art. 74 (2) ICC Statute does not merely describe the “accusatorial principle”, limiting the trial court’s inquiry to whatever conduct the prosecutor, in her discretion, chooses to include in the charging document. Since Art 74 (2) ICC Statute speaks of “facts and circumstances described in the charges”, it alludes to two further functions of the charging document, which are more clearly expressed in Regulation 52 (1) (b) of the Court’s Regulations: The charging document must, in a meaningful fashion, circumscribe the scope of the trial, and it must put the defendant on notice of the alleged facts to which he needs to direct his defense efforts. A loose description of a large time period and an area of thousands of  square kilometers plus a generic list of legal offense descriptions cannot fulfil either function. It is only when the relevant acts (or omissions) are described as to their approximate time and space that the defendant can reasonably prepare his response. As the history of the case at hand shows, an overly broad formulation of the charges also deprives both the pretrial and the trial chambers of a fixed target: the prosecutor may at any time add new incidents that fall under the broad purview of the loosely described original charges. One should, moreover, bear in mind that the charges and the ensuing findings of the court also describe the scope of the defendant’s protection against double jeopardy: if the charges were conceived as broadly as the Bemba Trial Chamber understood them to be, the court’s judgment (be it a conviction or an acquittal) would shield Mr. Bemba from responsibility for any atrocities his soldiers committed in the relevant time period even if new evidence showed that their crimes were much more extensive that assumed at the time of the trial. 

In the case at hand, the issue is complicated by the fact that Mr. Bemba was charged not with (direct or indirect) perpetration of the crimes in question, in accordance with Art. 25 (3) (a) ICC Statute, but with responsibility for his subordinates’ crimes as a commander. One might argue that the fault of the commander is his failure to take “all necessary and reasonable measures” to prevent or repress the subordinates’ offenses (Art. 28 (a) (ii) ICC Statute), hence the charges against him only need to describe the period of time in which he should have taken such measures. However, under Art. 28 the commander is not convicted simply of failing to exercise control over his troops (which might have been a sensible approach for the formulation of the Statute) but is held responsible for the very crimes that his subordinates committed. It is therefore imperative that the charges individualize each “underlying” crime, because the commander must be put in a position to raise a defense against each of them. He might do so, e.g., by showing that he was unable to prevent the subordinate from committing a specific offense, or that he would have had to take unreasonable measures in order to stop him from committing the crime in question.[9]

(3) Did Mr. Bemba take “all necessary and reasonable measures”?

With regard to the offenses subject to the Trial Chamber’s jurisdiction, the critical question is whether Mr. Bemba had taken all necessary and reasonable measures to prevent them and – if necessary – to submit them to the proper authorities with a view toward sanctioning the individuals responsible for the crimes. It was beyond dispute that Mr. Bemba had learned of some of these crimes and also had taken some preventive and repressive action in regard to them (installing commissions of inquiry, contacting UN agencies, seeking the cooperation of the CAR government, etc.). The critical issue thus was whether his activities were sufficient and genuine.

As to sufficiency, the two dissenters side with the Trial Chamber in finding that the efforts undertaken by Mr. Bemba to curb the misconduct of his troops was not all he could reasonably have done in that respect. The judges in the minority also fault their colleagues on methodological grounds: the majority, they claim, uncritically believe Mr. Bemba’s assertions about his lack of efficient means and fail to consider the totality of the evidence before the court, which – in their opinion – strongly suggests the defendant’s lack of genuine effort.[10]

The judges in the majority, by contrast, emphasize that the Rome Statute does not provide for a commander’s strict liability for offenses committed by his subordinates but explicitly limits the commander’s obligation to doing what is feasible and reasonable for the prevention and repression of crimes.[11] In accordance with the principle of individual criminal culpability, the majority claim, commanders can be held accountable only for the crimes of those they are commanding directly and whose conduct they can actually monitor.[12] The criterion of “reasonableness” implies further that a commander may take into consideration the impact of potential measures on ongoing or planned military operations and may choose the least disruptive measure as long as it can reasonably be expected to prevent or repress the crimes.[13] The majority also claim that the trial court must “specifically identify what a commander should have done in concreto” before he can be convicted based on his lack of activity.[14] In this respect, the number of crimes charged gains significance: since only slightly more than 20 offenses of Mr. Bemba’s soldiers were properly before the Trial Chamber, the majority judges argue, the demands on his activity must be proportionate to this number of crimes, not to a multitude of other offenses that DRC troops might have committed.[15]

In the arguments of both the majority and the minority of the Appeals Chamber, the motivation for Mr. Bemba’s activities play a major role. Both the Trial Chamber and the minority judges doubt the sincerity of Mr. Bemba’s efforts. The Trial Chamber saw his activities as motivated to a large extent by a desire to preserve the reputation of his troops and to rehabilitate their public image against public allegations of misconduct.[16] The majority of the Appeals Chamber, by contrast, emphasize that motives play only a limited role for the question of whether a commander undertook genuine efforts to prevent his subordinates’ crimes. The judges in the majority explain that a commander may well “discharge his duty to take ‘necessary and reasonable measures’ and in doing so accomplish multiple, additional or extraneous purposes, such as protecting the public image of his forces.”[17] They therefore see legal error in the Trial Chamber’s reliance on Mr. Bemba’s motives as grounds for discrediting the genuineness of his efforts.[18]

Another point of contention is the subjective standard for liability under Art. 28 (a) ICC Statute. The two dissenting judges write that it is not necessary, for the prosecution or the trial court, to make a clear distinction as to whether the accused “knew” or “should have known” about the (impending) offenses of his subordinates since Art. 28, in their opinion, provides for a “unitary standard” and hence the distinction “has no practical consequence”.[19] Judges Van den Wyngaert and Morrison, on the other hand, explain that the original charges, implying that Mr. Bemba “knew” of the offenses in question, could not subsequently be changed to the “should have known” standard since each alternative requires proof of different matters and hence can be defended against on different grounds.[20] Knowledge in the sense of Art. 28 of the ICC Statute “requires that the commander is aware that the subordinates are ‘about to’ commit crimes. This connotes a certain imminence and specificity going well beyond general concerns about the level of discipline of particular troops/units”.[21] The judges in the majority consequently assert that the evidence did not allow the Trial Chamber to find, beyond reasonable doubt, that the reports available to Mr. Bemba afforded him knowledge of imminent criminal conduct by his troops.[22]

Finally, the judges are of different minds with regard to the “as a result” formulation in Art. 28 (a) ICC Statute.[23] Although this dispute has no influence on the outcome of the case, it is interesting to note that the judges in the majority would in effect negate any significant substance of the “result” clause,[24] whereas the dissenting judges – deviating from their general tendency to give the Statute an extensive reading – would require proof of a high probability that the crimes in question would not have been committed if the commander had fulfilled his obligation to take reasonable preventive measures.[25]

I will comment only briefly on the interesting questions the judges raise concerning the interpretation of Art. 28 ICC Statute.

As to the fundamental issue of the extent of the commander’s obligation, the Statute appears to set an extremely high standard when it requires that the commander take “all necessary and reasonable measures … to prevent or repress” the commission of offenses by his soldiers. If the word “all” is taken literally, the very fact that the soldiers committed the crimes would indicate that the commander (who, according to Art. 28 (a) ICC Statute, exercises “effective command and control” over his subordinates) is indeed criminally liable, because his position of authority coupled with his knowledge of the imminence of offenses ought to have enabled him to interfere effectively, barring very exceptional circumstances. The judges in the majority correctly point out, however, that Art. 28 ICC Statute was not meant to confer strict liability on commanders, and that his options are to be assessed ex ante, not by hindsight.[26] I also agree with the majority’s position that the term “reasonable” in the description of the commander’s obligation significantly limits his liability. While I would not subscribe without qualification to the majority’s assertion that “commanders are allowed to make a cost/benefit analysis when deciding which measures to take”[27], I would still restrict the commander’s obligation to activities within his ordinary scope of authority.[28] Of course, preventing the commission of crimes is an important task, and the commander must carefully carry out that task because he has accepted a position of responsibility over a particularly dangerous instrument (i.e., a large group of heavily armed persons licensed to kill[29]). Yet, Art. 28 does not demand more of the commander than to “exercise control properly over [his] forces”, that is, to make use of the means conferred upon him in his capacity as a commander. I doubt, therefore, that the Trial Chamber in the instant case was correct in demanding that Mr. Bemba should have redesigned the campaign in CAR and/or withdrawn his troops from the CAR in order to protect the civilian population from harm.[30] These would have been strategic decisions which – though desirable – cannot in fairness be demanded under the threat of punishment in accordance with Art. 28 ICC Statute. In sum, I think that the majority view better reflects the limits of command responsibility. Whether Mr. Bemba in fact took all reasonable opportunities to prevent the crimes in question or to initiate an investigation of the soldiers involved in them is a matter of the available evidence, which cannot be properly assessed within the confines of this comment.

With respect to the question of whether Mr. Bemba made “genuine” efforts to prevent or repress the crimes in question, I do not think his motives are of relevance. It is the objective character of any measures the commander has taken that the court must assess; if these measures are sufficient under the reasonableness standard, it does not matter whether the commander took them in order to prevent crimes and to bring soldiers to justice, or to enhance his popularity, or even because he personally disliked the soldiers in question. If it was indeed Mr. Bemba’s motive – as the Trial Chamber asserted – “to protect the image” of his troops,[31] that motive may well have been in line with the legislative motive behind Art. 28 ICC Statute: Repression and sanctioning of soldiers’ crimes is demanded in order to preserve among the soldiers a law-abiding spirit, which will necessarily be reflected in the troops‘ image, both in public and within their own ranks. Although the majority judges correctly point out that a commander’s motives as such are irrelevant to Art. 28 charges, they may overstate the impact of the “motivation” factor on the Trial Chamber’s assessment of the overall sufficiency of Mr. Bemba’s measures.[32] It may thus be an over-reaction for the majority to dismiss the findings of the Trial Chamber in toto because of its alleged reliance on Mr. Bemba’s motivation.[33]    

Concerning the issue of the subjective element(s) of the commander’s responsibility under Art. 28 (a) (i) ICC Statute, the dissenters, in my opinion, take a wrong path in looking only at the identical end result of a conviction. It is true, of course, that a commander may be convicted if he “should have known” as well as if he “knew” of his soldiers’ impending or past offenses. But this does not mean that the charges are interchangeable and can simply be amended with regard to the evidence as it develops in the course of the trial. As the majority point out, it is of vital importance for the defendant to know whether he needs to counter the allegation that he was aware of the impending crime (which he can do by showing that no believable information of the soldiers’ plans reached him in time for taking effective counter-action) or the allegation (under the “should have known” alternative) that he should have become active in advance to learn about such plans. 

The divergence of opinions with regard to the “as a result” clause in this case is in part due to the lack of clarity of the text of Article 28. Whereas the French version does not seem to include any “result” requirement but only a conditional (“lorsqu’il”) connection between the commander’s failure to exercise control and his criminal liability,[34] the English version includes the words “as a result” but is grammatically ambivalent as to whether the resulting consequence of the lack of control is (a) the crime committed by the subordinate or (b) the commanders’ personal responsibility. Two arguments support the latter interpretation[35]: First, it leads to harmonious results under the English and the French versions; second, it avoids a logical contradiction with regard to the second alternative of Art. 28 (a) (ii), namely “to submit the matter to the competent authorities for investigation and prosecution”: Not submitting a case for investigation can hardly be regarded as the cause for antecedent crimes committed by the soldiers in question.[36] If one accepts this interpretation and views the commander’s criminal responsibility as the “result” of his lack of proper control over his forces (exemplified by the items listed in nos. (i) and (ii) of Art. 28 (a) ICC Statute), no causal element needs to be proved in these cases. If, on the other hand, one retains the more traditional interpretation, which demands a causal connection between the commander’s lack of proper control and the commission of the offenses by his troops,[37] then it makes sense to require no more than proof (beyond a reasonable doubt) of a high probability that the action expected of the commander would have prevented the commission of the crimes in question.[38] As the dissenters correctly point out, a hypothetical connection (What would have happened if the commander had taken the action reasonably expected of him?) can never be proved as a certainty.

(4) Standard of Review

Although the interpretation of Article 28 ICC Statute certainly played a role, the decision of the case eventually turned on questions of proof. The Appeals Chamber did not take any evidence but had before it the complete record of the trial. It seems that this record was not conclusive either in proving Mr. Bemba’s guilt or in showing his innocence of the crimes with which he was charged. The critical question then was whether the Appeals Chamber should defer to the findings of the Trial Chamber and accept its finding of guilt, or whether it should form its own judgment based on the evidence available in written and video form.

Both the majority and the minority of the Appeals Chamber delivered somewhat paradoxical opinions on this point. Judges Monageng and Hofmanski opted for a deferential attitude toward the Trial Chamber’s findings but nevertheless painstakingly examined the extensive record and used it to refute, one by one, Mr. Bemba’s objections to the finding of guilt. The judges in the majority, by contrast, emphasized the importance of determining the defendant’s guilt beyond a reasonable doubt on the basis of the trial record; but they did not undertake a thorough analysis of the evidence but instead issued a judgment of “not guilty” after having examined (and found wanting) the Trial Chamber’s judgment of conviction.

In their opinion, Judges Monageng and Hofmanski rely on the case law of the ICTY and in particular of the ICC Appeals Chamber in the Lubanga case for adopting a deferential standard of review on the appeals level with respect to alleged factual errors. Citing the Lubanga Appeals decision, they proclaim as the correct standard of review “whether a reasonable Trial Chamber could have been satisfied beyond reasonable doubt as to the finding in question. The Appeals Chamber will not assess the evidence de novo with a view to determining whether it would have reached the same factual finding.”[39] The main argument in favor of this standard is, of course, that only the trial judges are in a position to see and hear the witnesses and therefore have a more reliable and comprehensive access to the relevant evidence.[40]

The judges in the majority take a very different approach. In their opinion, “the Appeals Chamber must be satisfied that factual findings that are made beyond reasonable doubt are clear and unassailable, both in terms of evidence and rationale. Accordingly, when the Appeals Chamber is able to identify findings that can reasonably be called into doubt, it must overturn them.”[41] In essence, the judges in the majority place greater emphasis on the principle that guilt must be established beyond reasonable doubt, and they regard themselves as responsible for ascertaining that this principle is strictly applied and that any possible miscarriage of justice is avoided.[42] In support of their view, the judges cite two provisions of the ICC Statute: Art. 83 (1), which grants the Appeals Chamber, for the purposes of appeals proceedings, “all the powers of the Trial Chamber”;[43] and Art. 74 (5), which obliges the Trial Chamber to render a decision containing “a full and reasoned statement of the Trial Chamber’s findings on the evidence and conclusions”.[44] It is in fact the lack of a full and reasoned (and, one should add: convincing) statement of its finding of guilt which, for the judges in the majority, vitiates the Trial Chamber’s finding of guilt.[45]

It is difficult to fairly assess the relative strength of the arguments advanced by the opposing sides. Certainly, mere reliance on a principle of deference – which has no basis in the ICC Statute itself – appears unsatisfactory where the defendant may have been convicted without proof sufficient to overcome the presumption of innocence. The judges in the minority in fact go to great lengths in reviewing the Trial Chamber’s factual findings and in establishing their own view of the trial record. In that sense, the minority judges‘ actions (commendably) diverge from their theoretical stance.

One would indeed have expected that the judges in the majority undertake a similar task; but they do not.[46] For them, it seems sufficient to find that they have “reasonable doubts” about the Trial Chamber’s factual findings, because the Trial Chamber’s written judgment in itself appears to them to be inconsistent or unpersuasive. However, the majority‘s reliance in this regard on Art. 74 (5) ICC Statute concerning the obligation of the Trial Chamber to provide a full and reasoned statement seems beside the point. Tellingly, the majority struggle with explaining why a deficiency of the written judgment by itself should be sufficient grounds for overturning the verdict. Art. 83 (2) ICC Statute permits reversal of the Trial Chamber’s judment only if “the decision or sentence appealed from was materially affected by error of fact or law”. In order to justify reversal in Mr. Bemba’s case, the majority argue that a violation of Art. 74 (5) “has a material effect in terms of article 83 (2) of the Statute because it inhibits the parties from properly mounting an appeal in relation to the factual finding in question and prevents the Appeals Chamber from exercising its appellate review”.[47] That may well be true – but after the defendant has mounted an (obviously well-reasoned) appeal and after the Appeals Chamber has thoroughly considered the case, any such defect of the trial judgment is moot and can no longer be sufficient grounds for reversal. The dissenters’ criticism of the majority’s shortcut to reversal thus appears well founded.[48] I cannot quite suppress the thought that one motive for the judges in the majority to vote for Mr. Bemba’s acquittal may have been to sanction the Trial Chamber (and perhaps the Prosecutor) for what they regarded as shoddy work.[49]

But what are appeals judges supposed to do, under the ICC Statute, if they regard the Trial Chamber’s written reasons as insufficient to support a finding of guilt? The Statute in fact offers several options: According to Art. 83 (2), the Appeals Chamber may order a new trial, may remand a factual issue to the original Trial Chamber for it to determine the issue and to report back accordingly, or may itself call evidence to determine the issue. Reversal and retrial was indeed the resolution proposed by Judge Eboe-Osuji for the Bemba case.[50] Judges Van den Wyngaert and Morrison opposed this suggestion, mainly based on practical considerations, the most important of which was the length of detention that Mr. Bemba had already suffered and that would have continued for several more years in case of a retrial.[51] Judge Eboe-Osuji recognized that a stale mate within the Chamber would have ensued if he had insisted on his stance and therefore eventually agreed with the disposition suggested by his Colleagues in the majority.[52]

(5) Concluding Remarks

There are several perspectives that one can take on the outcome of Mr. Bemba’s case. The affair may be analyzed as a power struggle between the Trial Chamber and the Appeals Chamber, with the Appeals Chamber imposing its (majority) view on the other judges. One may also say that Mr. Bemba was lucky in escaping conviction due to excellent and inventive legal representation on appeal and a stale mate on the Appeals Chamber. But the uneasy resolution of the case also reveals certain deficiencies in the ICC Statute’s provisions on appeals proceedings. The Appeals Chamber certainly should not be in a position to simply replace the Trial Chamber’s factual findings by its own, without having thoroughly reviewed all the evidence beyond the reasons stated by the Trial Chamber. On the other hand, a retrial of the charges (be it before a Trial Chamber or an Appeals Chamber) in a typical ICC case would often be so cumbersome and time-consuming that it would impose disproportionate burdens on the accused who, after all, must still be treated as innocent. The rule of deference to the Trial Chamber’s factual findings might thus be a wise compromise; yet, this rule should not be applied too broadly and should leave enough room for the Appeals Chamber to correct potential miscarriages of justice.


[1] For details see Pros. v. Bemba, ICC-01/05-01/08-3343, Trial Chamber, Judgment of 21 March 2016, §§ 379 et seq.

[2] Hereinafter: Separate Opinion Van den Wyngaert & Morrison.

[3] Separate Opinion Van den Wyngaert & Morrison § 79.

[4] Pros. v. Bemba, Corrected Revised Second Amended Document Containing the Charges,  pp. 32-34.

[5] Pros. v. Bemba, Appeals Chamber, Judgment of 8 June 2018, Dissenting Opinion of Judge Sanji Mmasenono Monageng and Judge Piotr Hofmański (hereinafter: Dissenting Opinion) §§ 28, 32, 36.

[6] Pros. v. Bemba, Appeals Chamber, Judgment of 8 June 2018 (hereinafter: Majority Opinion) §§ 103, 104, 110, 115. The Majority also deem insufficient the formulation “including but not limited to” in regard to certain acts named as examples of soldiers’ offenses in the charging document .

[7] Majority Opinion § 119.

[8] Separate Opinion Van den Wyngaert and Morrison § 29.

[9] For this reason, I do not find persuasive the Minority’s argument that “in cases where command responsibility is alleged for mass crimes committed by the accused’s subordinates, the focus of the case will generally be the accused person’s ability and failure to exercise control properly, and the detail of the individual criminal acts alleged will generally be less material to the description of the charges under article 74 (2) of the Statute than in cases, for example, where the accused is alleged to have directly perpetrated those acts.” (Minority Opinion § 27).

[10] Minority Opinion §§ 52 et seq., 106.

[11] Majortiy Opinion §§ 168-169.

[12] Separate Opinion Van den Wyngaert and Morrison § 33.

[13] Majority Opinion § 170.

[14] Majority Opinion § 170.

[15] Majority Opinion § 183.

[16] See Pros. v. Bemba, ICC-01/05-01/08-3343, Trial Chamber, Judgment of 21 March 2016, § 728.

[17] Majority Opinion § 179.

[18] Majority Opinion § 191.

[19] Minority Opinion §§ 265-266.

[20] Separate Opinion Van den Wyngaert and Morrison §§ 38-40.

[21] Separate Opinion Van den Wyngaert and Morrison § 44.

[22] Separate Opinion Van den Wyngaert and Morrison § 50.

[23] “A military commander… shall be responsible for crimes… committed by forces under his or her effective command and control… as a result of his or her failure to exercise control properly over such forces, where …”

[24] Separate Opinion Van den Wyngaert and Morrison §§ 51-56.

[25] Minority Opinion §§ 321, 331, 335, 338, 339.

[26] Majority Opinion § 170.

[27] Majority Opinion § 170.

[28] Judge Eboe-Osuji in his Separate Opinion §§ 273-274 correctly points out that the obligation to take preventive measures may extend temporally to periods well in advance of the actual commission of crimes; however, there remains the subjective requirement that the commander must even at that time foresee or be able to foresse that without his action his troops will commit crimes.

[29] This concept is developed in the Separate Opinion Eboe-Osuji §§ 239, 274.

[30] Pros. v. Bemba, ICC-01/05-01/08-3343, Trial Chamber, Judgment of 21 March 2016, § 740.

[31] Pros. v. Bemba, ICC-01/05-01/08-3343, Trial Chamber, Judgment of 21 March 2016, § 728.

[32] The discussion of Mr. Bemba’s motivation is limited to one paragraph of the Trial Court’s lengthy opinion.

[33] See the criticism, along the same lines, in Minority Opinion §§ 72-73.

[34] Art. 28: […] Un chef militaire ou une personne faisant effectivement fonction de chef militaire est pénalement responsable des crimes relevant de la compétence de la Cour commis par des forces placées sous son commandement et son contrôle effectifs, ou sous son autorité et son contrôle effectifs, selon le cas, lorsqu’il ou elle n’a pas exercé le contrôle qui convenait sur ces forces dans les cas où […]

[35] This interpretation has been adopted in Separate Opinion Van den Wyngaert and Morrison § 56.

[36] Separate Opinion Van den Wyngaert and Morrison §§ 52-54. The Minority Opinion § 333 suggests that the problem could be solved by a requirement that “causation needs to be demonstrated in respect of subsequent crimes that were committed because of the failure to punish earlier crimes”; see also Separate Opinion Eboe-Osuji § 211. The problem with this approach is that the commander, under Art. 28 ICC Statute, is punishable for the “earlier” crime, not for the subsequent ones.

[37] As does the Minority Opinion §§ 331, 334, 335, citing the culpability principle and the principle of strict construction (Art. 22 (1) of the ICC Statute).

[38] Minority Opinion §§ 338, 339.

[39] Minority Opinion § 2, citing Pros. v. Lubanga, ICC-01/04-01/06, Appeals Chamber, Judgment of 1 Dec. 2014, § 27.

[40] Minority Opinion § 7.

[41] Majority Opinion §§ 3, 45.

[42] Majority Opinion § 40; Separate Opinion Eboe-Osuji §§ 9, 23, 52, 74, 79; Separate Opinion Van den Wyngaert and Morrison § 14.

[43] Separate Opinion Eboe-Osuji § 45. Judge Eboe-Osuji points out that no such provision existed in the Statutes of the ad hoc tribunals. However, it is unlikely that Art. 83 (1) ICC Statute is intended to grant the Appeals Chamber full authority to replace the Trial Chamber’s factual findings by its own (even without taking any evidence!). In light of the introductory words (“For the purposes of proceedings under article 81 and this article,…), a more probable interpretation of Art. 83 (1) would limit its impact to procedural matters, i.e., the authority to take the same procedural measures as the Trial Chamber.

[44] Majority Opinion §§ 49-50.

[45] This becomes quite clear from this statement in the Separate Opinion Van den Wyngaert and Morrison § 6: “A major concern we have is about the opacity of the Conviction Decision in terms of outlining the evidentiary basis for many of the findings. The decision is replete with cross-reference upon cross-reference and the reader is often left to speculate about which specific items of evidence the Trial Chamber relied upon for a particular finding.”

[46] The Minority Opinion § 18 is quite correct in asserting: “Even if the Majority were to be understood as requiring an assessment, under the standard of beyond reasonable doubt, of the accuracy of the Trial Chamber’s findings, it fails to do that as well.”

[47] Majority Opinion § 55. In a different context, the Majority Opinion (§ 50) argues that “the provision of reasons also enables the Appeals Chamber to clearly understand the factual and legal basis upon which the decision was taken and thereby properly exercise its appellate functions”.

[48] See Minority Opinion § 11.

[49] See what might be a revealing remark in the Separate Opinion Van den Wyngaert and Morrison § 73: “… we would not find it fair to give the Prosecutor a ‘second chance’ to prosecute this case, given the serious problems we have detected in the Prosecution case.”

[50] Separate Opinion Eboe-Osuji § 22; see also Minority Opinion § 53.

[51] Separate Opinion Van den Wyngaert and Morrison § 73.

[52] Separate Opinion Eboe-Osuji § 22.

How to Interpret Complicity in the ICC Statute

Thomas Weigend is a Professor of International, Comparative and German Criminal Law at the University of Cologne. At the risk of embarrassing him slightly, this is among the most insightful commentaries on complicity I have read in over four years of researching the topic.


My comment on the enigmatic words “for the purpose of facilitating the commission of such a crime” in Article 25 (3) (c) of the ICC Statute comes in two parts: First, is it good criminal policy for international criminal law (ICL) to require a higher degree of mens rea for convicting an assistant than for convicting a perpetrator of the same crime? Second, does Article 25 (3) (c) demand such a distinction by using the words “for the purpose”?

(1) As a general principle, it makes little sense to require a “thick” intention – however it is defined – for holding an assistant criminally responsible where a lesser degree of mens rea is sufficient for convicting the perpetrator of the same crime. The definition and sentence for any particular crime are devised with the perpetrator in mind. The allocation of responsibility to other persons, who have not “controlled” the commission of the offense or are otherwise further removed from the center of the criminal activity, is in some way accessorial to the perpetrator’s act. As the moon receives its light from the sun, an accomplice’s responsibility depends on an extension from the “natural”, primary responsibility of the perpetrator. Art. 25 (3) (c) clearly is based on this concept since this provision makes the commission or attempted commission of a crime by a perpetrator a prerequisite for convicting an assistant.  If that is so, it is counter-intuitive – and would require special justification – to add a specific mental requirement for convicting an assistant where the perpetrator will be held responsible on a less demanding standard.

This consideration is independent of whether one sees in Article 25 (3) (a) through (d) a “hierarchy” of modes of responsibility. Even if there is no such gradated system inherent in Art. 25 (3) and assisting is (or can be) of equal seriousness as perpetration, there is no good reason why an especially high degree of mens rea should need to be proved in order to convict an assistant. Some writers have proposed a “balancing” theory to justify this result: since the assistant does not singlehandedly complete the actus reus of the offense, they claim, her liability can be equal to that of the perpetrator only if the assistant’s mens rea is of a higher degree. But this calculus, to me, makes little sense.  If the assistant’s objective contribution is of lesser importance, then her sentence should reflect that fact. But the question whether the assistant desires the perpetration of the crime should have no influence on her punishment, because her “volition” does not increase the harm she causes or supports.

In ICL especially there is no good reason to require an “extra” degree of mens rea for convicting assistants. It makes little sense to exempt from responsibility those who know very well that the person whom they assist will make use of their contribution for committing a core crime but who have no direct personal interest in the perpetration of that crime. The commission of ICL crimes – contrary to many “ordinary” offenses – typically  requires the cooperation of many persons, who may all have different motives and goals. If ICL wishes to prevent such crimes it should not limit criminal liability to those who pursue a limited “purpose” and thus refrain from punishing all persons who consciously join their efforts to commit the offense.

(2) Does the wording of Art. 25 (3) (c) of the ICC Statute compel a different result? I don’t think so.

(a) Let us start with semantics. The “purpose” a person pursues describes his objective but not his motive. Therefore, an assistant under Art. 25 (3) (c) does not have to lend help because he wishes to bring about the offense. Nor does the commission of the offense have to be his sole purpose: Even if his main goal is to make money by selling arms, he may well act with the (secondary) purpose of facilitating the crime committed with those very weapons.

I also doubt that “purpose” necessarily coincides with a desire or with positive feelings about the (known) objective of one’s acts. For example, if a robber threatens to kill me unless I give him my wallet, and I comply with his demand, one can certainly say that I act with the purpose of satisfying his demand (so that he leaves me alone). But I certainly do not desire or even approve of his making off with my wallet. In sum, “purpose” describes one of the objectives of one’s act but does not say much about one’s attitude (of approval or disapproval) about that objective.

(b) This consideration leads to the general question about the role of “volition” in intent (or purpose). Some of the contributors to this blog seem to assume that – as James Stewart has put it – “the reference to purpose requires a volitional commitment to the consummated offense”. But what is the exact meaning of “volitional”, and why should it matter? If D shoots at V from close range, killing him, can D deny that he “willed” to kill V? Even if intention contains a “volitional” element, the actor’s volition will necessarily follow from his knowledge of the result that he is going to bring about. D may claim that he really did not like killing V (as in the robbery example above), and that he is sorry that he did – but still he “willed” V’s death, otherwise he would not have shot at him from close range. Contrary to  Flavio Noto, I would not say that it is a “fiction” that “certain knowledge about an undesired but anticipated side-effect is tantamount to a volitional commitment to that side-effect”. If the assailant in Flavio Noto’s airplane example knows that his rocket will kill all passengers, then it is his will to kill them all – although his motive for firing the rocket may have been his hate of one particular passenger. As in most other cases, the actor’s motive is irrelevant for his liability. His will flows from his act – if he had not willed the foreseen result he would not have acted.

This, by the way, is the hidden wisdom in the convoluted definition of intent and knowledge in Art. 30 of the ICC Statute: Normally, mens rea requires no more than a person’s awareness that a result will occur in the ordinary course of events. A further “volitional” element is necessary only where an offense definition requires that the perpetrator “intend” results beyond those brought about by the actus reus. If larceny is defined as taking someone else’s chattel with the intent of possessing it (as is the definition in § 242 German Penal Code), then it must be proved that the defendant “wished” to keep the chattel for himself – because that future development is not part of the actus reus of larceny. But assisting as defined in Art. 25 (3) (c) does not have this structure. The purpose of the assistant relates exactly to what she does: facilitate someone else’s (criminal) conduct. Therefore, the assistant’s volition is a necessary and undeniable consequence of her cognition.

(c) After having clarified what “purpose” may mean, we can turn to the question to what the assistant’s “purpose” must relate under Art. 25 (3) (c). The Statute speaks of “the purpose of facilitating the commission of such a crime”; the assistant’s purpose thus is not the crime but the facilitation. This means that the assistant’s objective must be to facilitate the act of the main perpetrator; but her will need not encompass the result of the perpetrator’s conduct. For example, if an arms trader sells weapons to a dictator, he will be punishable only if he does so with the purpose of facilitating the dictator’s use of armed force; but the fact that the armed force will be used against unarmed civilians and will therefore constitute a crime against humanity need not be the arms dealer’s “purpose” (although he needs to know about that particular use in order to be liable as an assistant under Art. 30 of the ICC Statute).

(d) In what I said so far, I assumed as true the widely shared assumption that the words “for the purpose” describe a special mental element of assisting under Art. 25 (3) (c). But there is a plausible alternative reading of these words, which has been spelled out by Antje Heyer in her excellent and extensive analysis of liability for aiding and abetting in ICL (published in 2013 in German under the title Grund und Grenze der Beihilfestrafbarkeit im Völkerstrafrecht, pp. 500-501; for a similar interpretation, see Katherine Gallagher, ‘Civil Litigation and Transnational Business’, 8 JICJ 745 at 765 (2008)). “For the purpose of facilitating the commission” can also be interpreted as an element of the actus reus of assisting: the assistant’s conduct must be specifically shaped in a way as to be of use to the perpetrator. Under this interpretation, conduct that is part of a person’s normal business would not qualify as assistance, because that conduct would not have the objective purpose of facilitating someone’s crime. If, for example, an arms trader sells weapons to a dictator at their regular price and under regular conditions, he would not be an assistant to crimes against humanity even if he is aware that such crimes will be committed using these weapons. But if the trader sells the weapons at a higher price because of an existing embargo, or if he sells weapons that have been specifically designed for killing civilians, he would be liable because this particular deal has been accommodated to serve the specific “purpose” of committing the crime. Under that interpretation, the regular mens rea requirements (as described in Art. 30) would apply – the arms dealer would only have to be aware of the specific elements that give the arms deal its “purpose”.