INTERNATIONAL CRIMINAL COURTS AND TRIBUNALS New court, same division: The Bemba case as an illustration of the continued confusion regarding the command responsibility doctrine

Prosecutor v. Jean-Pierre Bemba Gombo was the International Criminal Court ’ s first case directly addressing the command responsibility doctrine. As the first permanent international criminal court who can exercise jurisdiction over a majority of the world ’ s nation states, its interpretation of the doctrine was potentially an important development in international criminal law and an opportunity to affirm the legal responsibility of commanders for their subordinates ’ crimes. However, rather than providing a clear articulation of the doctrine and its scope, the Appeals Chamber was split. By a 3 – 2 majority, it reversed the Trial Chamber ’ s decision and the Appeals Chamber ’ s judges were sufficiently divided in their reasoning that they felt compelled to deliver separate opinions. A key disagreement within the doctrine is whether command responsibility is a mode of liability or a separate offence of dereliction of duty. This disagreement feeds into further contestation about the doctrine ’ s core elements, including the standard of fault necessary under its actus reus or mens rea elements. This article examines the judges ’ reasoning in Bemba to illustrate that, despite decades of jurisprudence and academic debate, there is still confusion on these foundational elements. Instead of being ‘ settled law ’ , the debate on command responsibility is still live. The article maintains that the current law supports a mode of liability interpretation but proposes that reclassifying the doctrine as a separate offence could resolve many of its tensions while observing the culpability principle, satisfying its justifications, and facilitating an adequately wide scope of accountability.


Introduction
The command responsibility doctrine allows commanders to be convicted of crimes committed by their subordinates when they fail to take all necessary and reasonable measures to prevent, repress or punish their subordinates' crimes. 1 In 2018, the Appeals Chamber of the International Criminal Court (ICC) acquitted Jean-Pierre Bemba Gombo (Bemba), who had been charged under command responsibility, of the war crimes of murder, rape, and pillaging and the crimes against humanity of murder and rape committed by his subordinates. 2 The Bemba case is significant because it is the ICC's first judgment on command responsibility. As the first permanent international criminal court whose jurisdiction is accepted by a large number of states, 3 its interpretation of the doctrine could have been an important step in clarifying the nature of command responsibility and defining its parameters. However, instead of providing that much-needed clarification, the Appeals Chamber's judgment reflects the division and disagreement that has long surrounded the doctrine.
The Appeals Chamber, by a 3-2 majority, reversed the Trial Chamber's judgment. 4 The appellate judges were also so divided that they felt compelled to express their varying interpretations in separate opinions. From the majority, Judges Van den Wyngaert and Morrison attached a Separate Opinion 5 , and Judge Eboe-Osuji attached a Concurring Separate Opinion. 6 Judges Monageng and Hofmański attached their Dissenting Opinion. 7 This article outlines how the judges disagreed over core issues, including the nature of the doctrine, the degree of knowledge or mens rea required, and whether causation is an element of the doctrine. All of these elements have long been disputed. The Bemba case demonstrates that, even after many decades of jurisprudence and academic debate, these disagreements have not been overcome.
A key question is whether the doctrine undermines the culpability principle and, if it does, whether that weakening of the principle is justified. The culpability principle is a core principle of many criminal law systems which requires that a person must have the requisite mens rea and actus reus to be convicted of a crime. When command responsibility is classed as a mode of liability, it generally places a strain on the culpability principle.
The division in the command responsibility doctrine reflects a dualism in international criminal law. International criminal courts are tasked with ending impunity for the most serious crimes of concern to the international community while also upholding the highest standards of criminal law and procedure. Perhaps in aiming to advance an interpretation that balances these roles, the judges in Bemba adopted varying positions on the core elements of the doctrine. While some of the interpretations of command responsibility as a mode of liability do not undermine the culpability principle, they set a high standard that narrows down the doctrine's scope and does not fulfil some of the key rationales advanced to justify it.
This article argues that reclassifying command responsibility as a separate offence would uphold the culpability principle, clarify the doctrine's foundational elements, and provide it with an adequate scope. It would also satisfy the key rationales advanced to justify weakening the culpability principle and bring the dual role of international criminal courts into greater harmony. I do not argue that the correct interpretation of the law, as it currently stands, is that it is a separate offence. Rather, the jurisprudence and the literal interpretation of the Rome Statute indicate that command responsibility is a mode of liability. Importantly, this article also does not question whether commanders who fail in their duty should be prosecuted but rather for what they should be prosecuted.
Given the uncertainties surrounding the doctrine, I start by outlining its more accepted parameters. In particular, Section 2 provides the Rome Statute definition of command responsibility and 3 There are currently 123 states that are party to the Rome Statute, see 'Status of Treaties: Rome Statute of the International Criminal Court', United Nations Treaty Collection, 17 July 1988, available at treaties.un.org/Pages/ViewDetails.aspx?src= TREATY&mtdsg_no=XVIII-10&chapter=18&clang=_en. 4 It is noteworthy that the Majority Judgment principally addresses 'all necessary and reasonable' measures and not other disputed elements of the command responsibility doctrine. See Majority Judgment, supra note 2, § 32. Cf.
Prosecutor v. Jean-Pierre Bemba Gombo, Judgment, ICC-01/05-01/08-T, Trial Chamber III, 21 March 2016 (Trial Judgment). 5 then contrasts the doctrine against traditional modes of liability in order to ascertain its outer limits by identifying what it is not. Section 3 outlines the confusion surrounding the nature of command responsibility and how disagreement on the doctrine's nature persists in the Bemba case. Section 4 considers how this disagreement extends to its elements, namely the mens rea and the actus reus elements. When classed as a mode of liability, some interpretations of command responsibility undermine the culpability principle but facilitate prosecution while others uphold the principle but unduly narrow down the doctrine's scope and hinder prosecution of serious crimes. Section 5 sets out key justifications advanced to weaken the culpability principle and extend the commander's liability. It notes that there are notable challenges to reclassifying command responsibility as a separate offence, but such a reclassification would satisfy these justifications without compromising the culpability principle and resolve many of the tensions within the doctrine.

What command responsibility is not
Article 28(a) of the Rome Statute sets out the command responsibility doctrine for military and de facto military commanders. 8 It states that the commander is 'criminally responsible for crimes : : : committed by forces under his or her effective command and control, or effective authority and control'. However, this criminal responsibility only arises if: (i) the crimes are committed 'as a result of his or her failure to exercise control properly over such forces'; (ii) the commander 'knew or, owing to the circumstances at the time, should have known that the forces were committing or about to commit' the crimes; and (iii) the commander did not 'take all necessary and reasonable measures within his or her power to prevent or repress their commission or to submit the matter to the competent authorities for investigation and prosecution'. 9 In short, Article 28 provides that commanders are liable if their subordinate commits a crime as a result of their failure to exercise proper control when they knew or should have known that the crime was being committed or about to be committed and they did not take all necessary and reasonable measures to prevent or repress the crime or to submit the matter to a competent authority.
Command responsibility allows holding commanders liable for their failure to effectively control their subordinates rather than for directly participating in their crimes. Indeed, if the commander participates in, contributes to, or assists in the commission of the crime, they would be liable either as principals or as accomplices under the traditional modes of liability pursuant to Article 25(3) of the Rome Statute. The traditional modes capture the conduct of a commander who orders or prompts the crime as well as a commander who does not initiate the crime but contributes to or aids and abets it. 10 If a commander intends to support the crime through their omission then they could also be responsible as an aider and abetter. 11 If a commander knows of a crime but fails to prevent it when under a duty to do so and has the material ability to prevent it, 8 Art. 28(b) addresses other superior-subordinate relationships, such as a civilian superior, which is outside the purview of this article. (ICTY Statute), Art. 7(3), which stipulates that the commander will be subject to 'criminal responsibility if he knew or had reason to know' that his subordinate committed or was about to commit the crime and did not 'take the necessary and reasonable measures' to prevent the crime or to punish the subordinate. 10 See D. Robinson then the mode of 'aiding and abetting by omission' potentially applies. 12 If a commander's omissions extend over a long period of time, their failure could constitute incitement and instigation as well as aiding and abetting in cases where the subordinates knew that the commander was aware of their continued criminal activities and the commander's omission contributed to the crimes. 13 Accordingly, the traditional modes of liability potentially capture a wide range of the commander's contributions, including their wilful omissions, intentional support or continued failures over a long period of time. Command responsibility provides for the commander's liability in other circumstances, namely, for failures falling short of those traditional forms of participation in or contributions to crime.
Command responsibility is also solely concerned with crimes committed by persons in a superior-subordinate relationship and with commanders' omissions, that is, their failure to prevent, repress or punish. Where the commander's omission could fall under both command responsibility or a traditional mode of liability, international criminal tribunals tend to convict under the traditional mode. 14 As such, command responsibility is 'a kind of default liability' for when the other modes of liability cannot be sufficiently established. 15 Command responsibility tends to work as a 'fall back' and entails a lower level of personal culpability than the traditional modes.

Confused nature of command responsibility
Command responsibility allows holding commanders liable for their subordinates' crimes, but the elements needed to establish the commanders' guilt do not relate to the underlying crimes but rather to the commanders' dereliction of duty to effectively supervise and control their subordinates. This divergence raises questions about the nature of command responsibility, namely whether the doctrine represents a mode of liability for the underlying offence committed by the subordinate or a separate offence of dereliction of duty by the superior. Are the commanders criminally liable for their subordinate's crimes, such as murder, rape, tortureor for an offence of failing to control, prevent or punish those crimes? Some maintain that command responsibility is a form of participation or mode of liability, such as accessory liability. 16 Indeed, this is the position favoured by numerous tribunals and finds basis in customary international law, as evidenced by the International Criminal Tribunal for the former Yugoslavia (ICTY) jurisprudence. 17 Yet, command responsibility is not consistent with traditional modes of liability. The mental standard may be lower than the one generally set for complicity and, under the ICTY legal framework, there is no requirement of a causal link  15 Ambos, ibid., at 1028. 16 See, for example, Robinson, supra note 10, at 58. Robinson asserts that the tribunals' 'applicable law seems to clearly establish command responsibility as a mode of liability'. See ibid., at 4-5. D. Robinson to the underlying crime. 18 The mens rea and actus reus standards may be different in command responsibility than under traditional modes of liability. 19 Others maintain that command responsibility is a separate offence; it is a crime of omission comprised of the commanders' failure to properly supervise and control their subordinates. 20 This position is supported by a number of scholars, especially in respect of commanders who failed to punish their subordinate's crime. 21 There is also some judicial support for the position that commanders are only responsible for their failure to fulfil their duty and not for the subordinates' crimes. 22 However, such judicial support is limited and generally originates from trial chambers or separate and dissenting opinions of individual judges and the underlying reasoning has faced academic challenge. 23 It is also difficult to class command responsibility as an independent offence when the commander's liability 'is strictly and necessarily dependent from the commission of the crime by the subordinate'. 24 Moreover, as expounded by Judge Eboe-Osuji in Bemba, Article 28 of the Rome Statute provides that commanders are responsible 'for crimes within the jurisdiction of the Court' and those crimes are genocide, crimes against humanity, war crimes, and the crime of aggression. Dereliction of duty is not a listed crime. 25 Also, bar a few limited exceptions, the actual practice of the international tribunals and courts, including in Bemba, has been to charge 18 Meloni, supra note 11, at 632. 19 Not all traditional modes of liability require the same subjective mens rea in order to establish culpability. In some common law systems, a person who participates in a crime can also be liable for additional incidental crimes committed by another member if the person could have foreseen the possibility of the additional incidental crime and participated in the plan/initial crime with that foresight. This position on the common law of complicity was recently upheld by the High Court in Australia, see Meloni, supra note 11, at 632. 25 Judge Eboe-Osuji Opinion, supra note 6, § 195. See also § § 194-202 on why the law, in its current format, would not be compatible with command responsibility as a dereliction of duty offence. Other support for this interpretation includes that command responsibility is not placed on the list of crimes in Art. 5 but is rather contained in the part on the 'general principles of criminal law', see Parts II and III Rome Statute respectively; Triffterer and Arnold, supra note 11, at 1087-8. See also Jackson, supra note 12, 114-15 on how statutory and judicial interpretation of the Rome Statute supports the mode of liability approach. commanders with, convict/acquit and sentence them for the same underlying crimes of their subordinates, as opposed to dereliction of duty. 26 This division between advocates of the mode of liability approach and the separate offence approach resonates with international criminal law's dual objectives: ending impunity for the most serious crimes of concern to the international community, on the one hand, and upholding the highest standard of criminal justice and procedures on the other. 27 These objectives do not have to but certainly can be in conflict. International courts seek to ensure that those culpable of serious crimes do not escape accountability due to, for example, the practical difficulties in prosecuting crimes. The rights of the accused must also be protected and the defendants must be afforded the highest standards of criminal justice. Those who advocate for classing command responsibility as a mode of liability often rely on arguments centred on ending impunity and ensuring that commanders do not evade justice. By contrast, advocates of the separate offence approach often rely on foundational principles of criminal law to support their position. 28 Perhaps in seeking to balance these objectives, scholars have also advanced interpretations of command responsibility that place it in the spectrum between separate offence and mode of liability. Some maintain that it is a form of participation that is akin to but distinct from the traditional modes of liability, 29 or a sui generis mode of liability or form of participation. 30 Indeed, classing command responsibility as sui generis was the predominant position at the ICTY. It is sui generis 'because it contemplates a species of criminal liability more comfortably explained by the theory of the commander's dereliction of duty than by the theory of vicarious liabilitythe latter entailing punishment of the commander for the crimes of subordinates'. 31 Several ICTY chambers also stated that the commanders do not bear the same responsibility as the subordinate. Rather, 'for' means that 'because of' the subordinate's crimes, the commanders are responsible for their failure to act. 32 Darryl Robinson highlights that what the ICTY sought to articulatenamely, that the commander is not a perpetrator and is not vicariously liable due to the commander-subordinate relationship but is responsible because of their fault with respect to the crimeis captured by the established mode of accessory liability. He argues that there is no need 'to fabricate an entire untested and vaguely-described conceptual category' by labelling command responsibility sui generis. 33 26 Robinson, supra note 10, at 35-6; Prosecutor v. Jean-Pierre Bemba Gombo, Judgment, ICC-01/05-01/08-T, Trial Chamber III, 21 March 2016, § § 2, 752 (Trial Judgment); Majority Judgment, supra note 2, § 116. 27 See, e.g., Rome Statute, supra note 1, Preamble for support that 'the most serious crimes of concern to the international community as a whole must not go unpunished' and Parts 3, 5-6 for high standards and procedures. 28 For a similar argument, regarding bodies of law in international criminal law see Sander, supra note 17, at 125-7. Sander points out that those who advocate for a mode of liability approach accentuate the public international law component and those who advocate for a separate offence approach accentuate the criminal law component. 29 Triffterer, supra note 11, at 1083-4 states that 'command responsibility is not merely an extending mode of the traditional responsibility for participation, though it comes close to facilitating or supporting crimes of subordinates in a way, similar to "otherwise assists in its commission or its attempted commission"'; Triffterer and Arnold, supra note 11, at 1105, stating that command responsibility 'does not substitute, but supplements all forms of participation as listed in Article 25(3)'. 30 Meloni, supra note 11, at 632. 31 Judge Eboe-Osuji Opinion, supra note 6, § 188 referring to the ad hoc tribunals' jurisprudence.  33 Robinson, supra note 10, at 38-9. He does, however, note that command responsible is 'distinct' as it only relates to superior-subordinate relationships and features a more lenient mental element; he is mainly arguing against the lack of a contribution requirement. See ibid., at 36-9. See  Others argue that command responsibility is a sui generis liability by omission, 34 or that it is a sui generis crime, 35 or that it is a mixture of a sui generis form of participation and a sui generis criminal act depending on the circumstances. 36 Similarly, command responsibility has been explained as 'a responsibility for-the-act-but-not-for-the-act', as 'sometimes-mode-sometimesoffence' and as 'neither-mode-nor-offence'. 37 The picture is further complicated in view of the arguments that, although it is characterized as a sui generis form of liability by the tribunals, it is de facto equated to a separate offence in the practice of the tribunals, 38 or that command responsibility is an imputed liability, vicarious liability or criminal negligence. 39 In Bemba most of the judgments or individual opinions did not address in detail the doctrine's nature. 40 Where it was addressed, the judgments and opinions neither resolved the debate nor provided clarity. The Trial Chamber followed the prevailing ICTY jurisprudence and classed command responsibility as a sui generis mode of liability. 41 Judge Steiner, however, preferred the term 'additional' as opposed to 'sui generis'. 42 The Trial Chamber also adopted similar rhetoric regarding whether the commander shared the same responsibility as the subordinate. It stated that commanders are 'criminally responsible for crimes : : : committed by [their subordinates]' but 'the responsibility of a commander : : : is different from that of a person who "commits" a crime within the jurisdiction of the Court'. 43 Unfortunately, the Chamber did not elaborate on how exactly the responsibility is different when the commander is found responsible 'for' the same crime as the principal perpetrator and did not engage deeply with the debate on the doctrine's nature. On appeal, Judge Eboe-Osuji disagreed with the Trial Chamber and stated that command responsibility was not sui generis. 44 Instead, he classed it as accomplice liability and specified that it is a 'special provision of accomplice liability' to the general provision of accomplice liability outlined in Article 25(3)(c) of the Rome Statute. 45 In short, the nature of command responsibility has been the subject of considerable debate wherein varying positions have been advanced. The prevailing ICTY jurisprudence which favoured interpreting command responsibility as a sui generis mode of liability was subjected  35 See, e.g., Triffterer and Arnold, supra note 11, at 1095, 1088, 1105, reflecting Arnold's opinion. Triffterer states that command responsibility is not a crimen sui generis, see Triffterer, supra note 11, at 1083. 36 See, e.g., M. L. Nybondas, Command Responsibility and Its Applicability to Civilian Superiors (2010), 133-9. See also R. Arnold, 'Maria L. Nybondas, Command Responsibility and Its Applicability to Civilian Superiors', (2013) 11 JICJ 931, at 945-7; Triffterer and Arnold, supra note 11, at 1087. 37 Robinson, supra note 10, at 1, 3.  Galand, supra note 33, at 684-5 highlights that the Bemba Appeals Chamber neither clarifies the nature of the commander's responsibility nor addresses its elements, such as mens rea, and instead these are 'touched upon' in the Separate and Concurring opinions. 41 Trial Judgment, supra note 26, § § 171, 173, 174. 42 Ibid., at note 388. 43 Ibid., §173. 44 Judge Eboe-Osuji Opinion, supra note 6, § 198. 45 Ibid., § § 217-218. Judge Eboe-Osuji uses the endangerment theory to explain the commander's criminal responsibility under the doctrine. See ibid., § § 251, 232-53, 267.
to scholarly criticism and a number of scholars endorsed the interpretation of the doctrine as a separate offence of dereliction of duty. Rather than resolving this debate, the Bemba case reflected the confusion. Within the same Court, the judges were either silent on the issue or contradictory positions were advanced, evincing that the issue of the doctrine's nature is still live. This is not only of theoretical concern but has serious practical repercussions, in particular it can determine whether a person is found liable of egregious crimes or not. It could also lead to serious inconsistencies within international criminal law and undermine its authoritativeness.

Stretching the culpability principle under the mode of liability approach
The culpability principle provides that to be criminally liable a person must have personally engaged or participated in the crime in some way. 46 This is a core principle for liberal criminal justice systems, which international criminal law purports to be, 47 and is well-accepted in the criminal law of all major national legal systems. 48 Judges Monageng and Hofmański recognized the importance of this principle to international criminal law in Bemba. They stated that the commander's criminal liability for their subordinate's crime 'is only justified and indeed justifiable if there is a personal nexus between the crime and the superior'. 49 Culpability or personal guilt has two elements. It requires the objective element of 'a personal connection to the crime' and the subjective element of 'a blameworthy mental state'. 50 Generally, a person should only be liable if they have contributed to the crime and have the requisite knowledge or intent. As a mode of liability, command responsibility potentially undermines both in that the commander is punished 'for [his] subordinates' crimes, when he neither participated in the actus reus nor shared the mens rea of the crimes in the primary sense of desiring it'. 51

Mental element: A blameworthy mental state
Traditional modes of liability generally demand intent and actual knowledge or some level of subjective awareness. 52 While many agree that command responsibility requires 'something lower' than the traditional modes -Judges Monageng and Hofmański in Bemba stated that the mental standard is 'significantly lower' -53 how much lower is contested. Some of the positions advanced stretch the culpability principle further than others. Whether intent is required and, if so, which degree of intent, the standard of knowledge required, and what knowledge is required all impact on the culpability principle. 46 See, e.g., Prosecutor v. Duško Tadić, Judgement, IT-94-1-A, Appeals Chamber, 15 July 1999, §186; Robinson, supra note 10, at 12; Sander, supra note 17, at 123. 47 Robinson, ibid., at 4, 14. 48 Triffterer, supra note 11, at 1060; M. Damaška, 'The Shadow Side of Command Responsibility', (2001) 49 American Journal of Comparative Law 455, at 464. There are, however, differences between nations in respect of their adherence to the culpability principle and the extent to and circumstances in which states will allow the principle to be compromised. See Damaška ibid., at 464-5. 49 Judges Monageng and Hofmański Opinion, supra note 7, § 334. 50 Robinson, supra note 10, at 12.  52 See, e.g., Rome Statute supra note 1, Art. 30(1), which states that: '[u]nless otherwise provided, a person shall be criminally responsible and liable for punishment for a crime within the jurisdiction of the Court only if the material elements are committed with intent and knowledge'. The provision of 'unless otherwise provided' means that the lower standard set out for command responsibility in Art. 28 is an exception to this general rule. 53 Judges Monageng and Hofmański Opinion, supra note 7, § 331.
The mens rea that the commander must be proven to possess has been variously argued to amount to wilfulness or acquiescence to mere negligence. 54 These represent vastly different standards and burdens for the prosecutor to satisfy before a commander can be found liable. The different Bemba judges' reasoning, often provided in dicta, reflects the intractable uncertainty. 55 At one extreme, Judge Eboe-Osuji set a very high standard and required that commanders wilfully fail to exercise effective control in a manner that 'truly amounts to connivance in or condonation of' the subordinate's crimes before they are responsible for their subordinate's crimes. 56 At the other extreme, the ICC Pre-Trial Chamber stated that commanders are criminally liable if they have been 'merely negligent in failing to acquire knowledge' of their subordinates' crimes. 57 The 'negligence' position is supported by the low knowledge standard under the doctrine. Command responsibility imposes liability when the commander had actual knowledge ('knew') but also when the commander 'should have known' of the crimes. 58 On appeal in Bemba, Judges Van den Wyngaert and Morrison argued that the 'should have known' standard captures situations where the commander has an 'awareness that something is going on without having sufficiently clear and dependable information as to what is happening/has happened, when it is going to happen/has happened, or who is/was involved'. 59 A subjective 'awareness that something is going on' is a notably lower mental state standard than acquiescence, connivance or condonation or, indeed, the standard generally required in traditional modes of liability.
Under the Pre-Trial Chamber interpretation of 'should have known', commanders may not even need a subjective awareness that something is going on and instead the standard required 'an active duty : : : to take the necessary measures to secure knowledge of the conduct of his troops and to inquire, regardless of the availability of information at the time on the commission of the crime'. 60 Under this construction, negligence in failing to act when aware that something may be happening would not need to be established; negligence in failing to actively seek information and monitor, even when there is no awareness of wrongdoing, would suffice.
The mental standard is also potentially lowered by interpretations on what knowledge the commander needs to possess. The Trial Chamber in Bemba stated that the commander does not need to know all the details of each crime or the specific identities of the perpetrators. In addition, knowledge of the commission of the crime was held to 'necessarily [imply] knowledge of the requisite contextual elements which qualify the conduct as a war crime or a crime against humanity'. 61 The contextual elements convert particular crimes into war crimes or crimes against humanity and, accordingly, knowledge of these elements is central to the conviction of the principal perpetrator. Several international crimes also require that the principal perpetrator had not 54 Cf. Mettraux, supra note 12, at 78 (commander's failure must amount 'to a form of "personal neglect amounting to a wanton, immoral disregard of the action of his subordinates amounting to acquiescence"', citing the High Command case, and see 63-4, 73-4); Triffterer and Arnold, supra note 11, at 1099 ('Mere knowledge, or failure to acquire knowledge where this would have been required by the circumstances, is per se enough. This kind of failure : : : may constitute : : : unconscious negligence'). 55 For example, the majority in the Appeals Chamber acquitted Bemba on the 'all necessary and reasonable measures' element rather than the 'mens rea' or lack of 'causation' elements. See Majority Judgment, supra note 2, § 32. 56 Judge Eboe-Osuji Opinion, supra note 6, § § 10-12, 199-202, 268-269. Galand maintains that the standard advocated by Judge Eboe-Osuji may not be as high as it appears at first blush and that when read in conjunction with the endangerment theory's reasoning, 'wilful' would include failing to implement preventive measures, such as establishing and monitoring a reporting system: see Galand, supra note 33, at 694. only knowledge and intent, but rather a special intent. For example, genocide, torture, and pillaging all require that the person participates in the crime for a specific purpose, such as, to destroy a group, to extract information or impose punishment, or to take goods for private or personal use respectively. It is this specific intent that 'elevates' murder to genocide, beatings to torture, and theft to pillage. The requirement of this higher mental state reflects the graver character of the crime. If command responsibility allows holding commanders liable as a principal, then it entails the much lower standard of knowledgeor even hypothetical knowledge under 'should have known'than special intent. 62 Judges Van den Wyngaert and Morrison's position is that a subjective 'awareness that something is going on' satisfies the 'should have known' standard, which is notably lower than wilful, connivance or condonation. As opposed to personal intent or sharing the principal perpetrator's intent, the prosecution would only need to establish knowledge or even hypothetical knowledge under 'should have known'. 63 This lowered mental state has led many scholars to hold that the command responsibility doctrine allows the negligent commander to be convicted of the same crimes as their subordinates. 64 If command responsibility is a form of principal liability, it converts a negligent omission into an intentional criminality. 65 It is 'a crime of intent by negligence'. 66 It may even be a crime of special intent by negligence. It could be genocide, torture, or pillage by negligence.
However, the potential divergence between culpability under the traditional modes of liability and under command responsibility depends on whether the doctrine is classed as a form of accessory liability. Some, including Judge Eboe-Osuji in the Bemba appeal, maintain that command responsibility is a mode of accessory liability whereas others argue that it is not. 67 If this doctrine is a mode of liability akin to principal liability, then there is potentially a substantial culpability gap between it and other modes of principal liability. On the other hand, if the doctrine is a mode of liability akin to accessory liability, then the more appropriate comparable standards for mens rea or actus reus are those pertaining to accessory rather than principal liability. 68 This would narrow the gap between the culpability standard under the traditional modes and the doctrine because lower standards of knowledge, intent and contribution are seen in several traditional modes of accessory liability. 69 Under accessory liability, the person does not need to share the same mens rea with the principal. Accessories do not necessarily have to have the same (special) intent or the same knowledge of the details of the crime as the principal. 70 Nevertheless, negligence is often 62 See Stewart, supra note 21, at 196 (referring to knowledge in complicity being used to satisfy special intent). Cf. Robinson, supra note 16, at 652-7 who argues that command responsibility is a mode of liability akin to accessory liability and, accordingly, the more appropriate comparable standards for mens rea or actus reus is accessory liability rather than the principal perpetrator. 63 considered too low a standard for complicity, even for accessory liability. 71 The interpretations of command responsibility that favour lower standards of intent and knowledge, including the 'subjective awareness that something is going on', combine to place the doctrine at the weaker end of the culpability principle. It is questionable whether weakening the principle to this extent is necessary and justified in order for the doctrine to be effective.

Physical element: 'Personal connection' to the underlying crime
Traditional modes of liability generally require that a person participate in, contribute to, or have an effect on the commission of the crime. 72 While the importance of this 'personal connection' to upholding the culpability principle is well-recognized, 73 what the commander's personal connection means under the command responsibility doctrine is disputed. There are varying views on whether causality must be established. 74 If it must, what is the degree or the extent of contribution needed? There is even disagreement on whether the subordinate needs to have completed the underlying crime or whether merely attempting the crime is sufficient. 75 One of the most contentious issues arises with respect to the commander who fails to punish their subordinate's crimes. Punishing a crime requires the commander to act after the crime has already been committed and, accordingly, the commander's failure to do so cannot retroactively cause or contribute to that particular crime. The commander's failure to punish may contribute to the commission of future crimes, for example, by reducing the perceived risk of punishment, failing to impart the appropriate values or even condoning or encouraging the crimes. 76 However, when assessing causality under command responsibility, the ICTY did not restrict its interpretation to failures to punish that contribute to subsequent crimes. Instead, in numerous judgments, the ICTY held that causality is not required under the doctrine. 77 Indeed, in Čelebići the Trial Chamber used the inability of the commander's failure to punish past crimes to contribute to their commission to justify the lack of a causality requirement: no : : : causal link can possibly exist between an offence committed by a subordinate and the subsequent failure of a superior to punish the perpetrator of that same offence. The very existence of the principle of superior responsibility for failure to punish, therefore : : : demonstrates the absence of a requirement of causality as a separate element of the doctrine of superior responsibility. 78  73 See, e.g., Judges Monageng and Hofmański Opinion, supra note 7, § § 333-334; Trial Judgment, supra note 26, §211. 74 See, for example, Trechsel, supra note 34, at 30 (no causal link needed); Robinson, supra note 10, at 57-8 (causal link needed). 75 Triffterer, supra note 11, at 1084; Triffterer and Arnold, supra note 11, at 1088-9, 1100, 1104; Mettraux, supra note 12, at 79-80 (completed). Again, rather than providing unity, Bemba sets out varying and conflicting positions. While some judges followed the ICTY approach, others did not. The Trial Chamber held that there must be a causal link between the commander's omission and the crime committed. 79 This reasoning aligns with a literal reading of the Rome Statute, which requires that the subordinates' crimes are committed 'as a result' of the commander's failure to exercise proper control. 80 By contrast, the ICTY Statute does not contain an explicit reference to a causality element. 81 Judges Monageng and Hofmański agreed that causation was necessary under the Rome Statute's construction but distinguished between the different duties of the commanders. They held that the causality element only related to the commander's duty to prevent future crimes. They reasoned that the commanders' failure in their duties to punish or repress crimes cannot retroactively cause past crimes. The failure to punish or repress crimes could cause subsequent crimes though. 82 In a similar vein, Judge Eboe-Osuji stated that it must be shown that the commander's failure caused the crime and that a failure to punish would not contribute to a past crime, such as the first or isolated crimes, but the failure could more readily, though not necessarily, be shown to have resulted in the subordinates' subsequent crimes. 83 However, using the same reasoning of the illogicality of a failure to punish causing a past crime and arguing that while the failure could cause subsequent crimes, this would only be 'in very exceptional circumstances', Judges Van den Wyngaert and Morrison reached a very different conclusion and held that there is no causality requirement for any of the commander's duties. 84 They stated that such an interpretation would lead to the 'absurd result' of generally making it 'impossible' to apply the duty to punish. 85 Even if there is an agreement that causality is requiredwhich appears to be the leaning of the ICC judges, although there is less consensus on when causality must be shownthere is renewed disagreement over the strength or degree of causation needed. For example, does there need to be a 'substantial connection' between the commander's omission and the crime or is an 'increased risk' that the crime would occur sufficient? 86 Again, this uncertainty was reflected in the separate opinions in Bemba. The Trial Chamber shied away from clarifying the degree of causation required. Instead, they merely asserted that it is a lower standard than the 'but for' test. 87 On appeal, Judges Monageng and Hofmański held that the appropriate test was one of 'high probability'. That is, there must be a high probability that if the commander had fulfilled their duty, the crime would have been prevented or would not have been committed in the manner it was committed. 88 However, Judges Van den Wyngaert and Morrison explicitly rejected this test. They stated that the test does not withstand critical analysis because the commander's duty is to reduce the risk of a crime occurring and a failure to reduce a risk cannot be said to cause 'the manifestation of said risk'. 89 Judge Eboe-Osuji seemed to favour a test of whether the contribution was a 'significant and operating cause' where 'significant' should be interpreted as a contribution that is 'more than 79 Trial Judgment, supra note 26, § § 210-213. 80 negligible or not to be so minute that it will be ignored under the 'de minimis' principle. 90 These represent vastly different standards stretching the continuum from a 'more than negligible' contribution all the way to a 'high probability' that the omission contributed to the crime.

The combined effect: Dispersed standards and a weakened culpability principle
The variances between the ICTY and ICC's stances can arguably be explained by the differences in the texts of the ICTY Statute and the Rome Statute. 91 However, such differences fail to fully explain the divergence of positions and confusion on key elements. When applying the same Statute in Bemba, the judges reached starkly different conclusions: the interpretive spectrum stretched from negligence to wilfulness and from there being no causation requirement to a requirement of a 'high probability' causation. The Bemba case illustrates not only that the doctrine is far from 'settled law', but also that the persisting division does not stem solely from discrepancies in the text of the Statutes. Instead, it indicates a deeper uncertainty and disagreement regarding the foundations of the doctrine itself.
At one end of this spectrum, command responsibility could be construed from the various positions in Bemba as requiring a wilful failure amounting to connivance or condonation and a 'high probability' that the omission contributed to the crime. While this interpretation would not undermine the culpability principle as it requires a guilty mind and a personal connection to the crime, it imposes 'an exacting burden upon the Prosecutor'. 92 This high threshold would align with the ICC's role of upholding high criminal justice standards. However, it could also limit the doctrine's effectiveness as blameworthy commanders may avoid liability. 93 This higher standard of wilful connivance and clear contribution may also result in the commander's omission being more readily captured by the traditional modes of liability and reduce the need for, or eclipse the purview of, command responsibility. This narrowing of the doctrine's scope may affect the ICC's ability to fulfil its role of ending impunity for the most serious crimes, at least under this doctrine.
At the other end of the spectrum, the varying positions in Bemba could also mean that the command responsibility requirements are satisfied if a commander failed to punish a subordinate for an isolated crime after the commander negligently failed to actively seek information, even though the commander was unaware of the crime and neither caused nor contributed to the crime. This standard would set a very low burden for the prosecutor, alleviate concerns of blameworthy high-ranking commanders evading liability, and arguably assist prosecution and punishment for the most serious crimes. However, this standard would also significantly undermine the culpability principle. The above represents an extreme interpretation but less extreme ones could also have a negative effect. This is because the principle is being potentially undermined on several fronts. It is notably compromised when negligence is substituted for intent and knowledge, converting negligence into a crime of intent or special intent. The culpability principle is further compromised by interpretations that negligently failing to secure knowledge irrespective of awareness of wrongdoing at the time is sufficient. It is weakened further by interpretations that liability can be imposed without establishing knowledge of specific details of the crime or its contextual elements. While each of these interpretations dilutes the culpability principle, the combined effect can be a substantial erosion of the principle.
The weakening of the culpability principle contrasts with the severity of the crime, which is particularly high. If a central tenet for assuring just attribution of liability is eroded when determining whether a person is guilty of war crimes, crimes against humanity or genocide, this calls into question whether the courts are upholding the highest standard of liberal criminal justice.

A different means to the same end
While it is not uncommon in municipal laws for the culpability principle to be 'stretched', the imposition of liability on a person for someone else's actions is generally only for lesser offences that do not entail 'serious moral condemnation' or strong stigmatization. 94 War crimes, crimes against humanity, and genocidefor which the commander can be held liable for under command responsibilityrightly carry strong moral condemnation and stigmatization and do not fall within the category of offences generally exempted from the culpability principle. Reclassifying command responsibility as a separate offence would allow upholding this principle while providing benefits in terms of conceptual clarity and effectiveness.

Justifications to extend the commander's culpability
Justifications advanced to rationalize the weakening of the culpability principle under command responsibility include that the commanders have voluntarily assumed the responsibilities of command and have the power of control over their subordinates. Armed conflict historically involves serious crimes and there are numerous factors that increase the risk of unlawful violence. Soldiers are trained in the use of violence, desensitized to it, and provided with weapons; the enemy can be dehumanized and military culture can promote obedience and group loyalty. These factors can combine to distort inhibitions against violence and killing. 95 Commanders have important duties to monitor and restrain their subordinates and the failure to fulfil them can lead to irreparable harm. 96 As persons in positions of authority and with control over their subordinates, commanders are best positioned to prevent and repress crimes and to ensure respect for international humanitarian law. 97 In recognition of this, international law imposes legal duties on belligerent states and their commanders to prevent and suppress their subordinate's crimes. 98 The extension of the commander's liability may also be justified by its deterrent value. The commanders' knowledge that their failure to properly control their subordinates could lead to their own conviction should motivate them to be vigilant. This is especially so given that the determination of their guilt will often be an ex post assessment of what they could have done by judges with little to no experience in the conditions and pressures of warfare. 99 Making commanders liable also potentially has a deterrent effect on their subordinates, who could not count on avoiding punishment for their own crimes when their commanders are called to account. 100 Setting a lower threshold for liability achieves these objectives because an actual knowledge standard could enable commanders, especially high-ranking ones, to use the 'hierarchical distance' between them and their subordinates to conceal their involvement or acquiescence or to falsely claim that they were unaware of the crimes. A negligence standard would more readily capture the 94 Damaška, supra note 48, at 464-5, 470. 95 See, e.g., Robinson, supra note 16, at 658-9. 96 Trial Judgment, supra note 26, § 172.  99 Damaška, supra note 48, at 473. 100 Triffterer, supra note 11, at 1069. behaviour of a commander who wilfully disregards information or knowingly participates in the crime where this knowledge cannot be proven. It sedates the fear that those responsible will 'slip between the cracks' and escape liability. 101 These justifications align with and reinforce the international criminal courts' role of ending impunity.
However, in a liberal criminal justice system, a person should not be held criminally liable solely because their conviction could serve a social benefit, such as deterrence. Such utilitarian grounds alone are not sufficient. There also needs to be deontological grounds that substantiate that the conviction is 'deserved'. 102 Convicting commanders for serious crimes under a doctrine that compromises the culpability principle raises concerns over whether the conviction imposed is 'deserved'.
Weakening the culpability principle may also have the inverse effect on its intended objective of ending impunity. Judges Van den Wyngaert and Morrison in Bemba use the culpability principle to warn against the 'mind-set' that assumes the high-ranking commanders to be criminally liable. They argue that it is the primary responsibility of low-ranking commanders to prevent, repress or refer the subordinate to the appropriate authority. While command responsibility can apply to high-ranking commanders, they will generally only be criminally liable for a low-ranking soldier's crime when it is shown that they did not monitor the soldier's superiors properly. 103 The Bemba Appeals Chamber majority warned against the 'very real risk' of evaluating what actions the commander should have taken with the benefit of hindsight. When assessing what is meant by 'all necessary and reasonable measures', 'all' should not be equated to the commander being required to take 'every single conceivable measure within his or her arsenal, irrespective of considerations of proportionality and feasibility'. 104 Instead, as long as the chosen measure is reasonably likely to prevent or repress the crimes, the commander is permitted to make a cost/benefit analysis and weigh up the impact of different measures on military operation when determining which measures to choose. 105 These interpretations advanced by the judges in Bemba potentially limit conflict with the culpability principle but they also restrict the doctrine's scope. Limiting the duty to prevent, repress, and report crimes to the purview of lower-ranking commanders or incorporating a cost/benefit analysis into the 'all necessary and reasonable measures' test arguably runs counter to the doctrine's original justifications. Such restrictions could undermine its ability to deter high-ranking commanders from using hierarchical distance to avoid culpability. They could weaken the imperative for those who have voluntarily assumed leadership duties to be vigilant and monitor their subordinates appropriately to prevent the most serious crimes. In turn, this could lead to fewer high-ranking commanders being held liable and impact on the Court's ability to end impunity.

A path forward: Reclassifying command responsibility as a separate offence
Reclassifying the doctrine as a separate offence could allow for an adequate scope of commanders' responsibility without undermining the culpability principle. It would help ensure that all commanders are accountable for their failures in a way tailored to their own degree of culpability. Focusing on the commander's omission provides a stronger deontological basis and ensures that the conviction is 'deserved'. Commanders should be convicted for their own actions or omissions and not those of other persons. For example, a negligent commander or the commander who fails to punish their subordinates is guilty and deserving of punishment for their failures and omissions. 106 Commanders assume the duty to supervise personnel trained in violence where there is a substantial risk that violations of the law could be committed against the most vulnerable people. Rather than a lesser offence, a commander's dereliction of this duty should be a serious offence in and of itself. It deserves stigmatization and punishment in its own right.
To an extent, the drafters of the Rome Statute recognized this by focusing on dereliction of duty when articulating the doctrine's elements even though, as the Statute is currently formulated, the doctrine's nature can most appropriately be classed as a mode of liability. Yet, focusing on the commanders' failure to fulfil their duty but convicting them for the different underlying crime committed by their subordinate creates a divergence between the elements of the crime established and the crime for which they are convicted. The stipulated elementsa failure to prevent or repress crimes or to submit to competent authorities the crimes of subordinates under their effective controlcould form a separate offence. That is, the elements already stipulated in the statutes could form the basis of the doctrine as a separate offence. This approach could also align better with command responsibility's origins in the 'responsible command' principle. Responsible command represents the general duty of commanders to control their subordinates, including by ensuring that they are properly organized, disciplined and compliant with international humanitarian law and standards. 107 Command responsibility is a broad doctrine that spans from the lowest to highest ranks of commanders and from negligent omission to condonation. Determining its exact parameters, content and elements would require an in-depth analysis beyond the scope of this inquiry. Instead, it focuses on the preliminary step of demonstrating that the ongoing uncertainties are linked to the doctrine's classification as a mode of liability and the encroachment of the culpability principle. Reclassifying command responsibility as a separate offence would, prima facie, resolve these uncertainties while providing this doctrine with a stronger deontological grounding and upholding the culpability principle.
A separate offence would require a graduated approach to determining the extent of the commander's failure and a corresponding graduation of the punishment to be imposed. This graduated approach could capture the full range of the commander's liability, for example from the commander who allows minor lapses in discipline to the commander who negligently fails to exercise sufficient control where no crime occurs to the commander who negligently fails to prevent or punish widespread and systematic crimes to the commander who recklessly fails to prevent or punish widespread and systematic crimes and so forth. The international tribunals already tend to rely on the traditional modes of liability for commanders who contribute to or participate in the crimes.
The proposed approach would mean that rather than the commanders' conviction being reliant on the crime of another, the determination of whether a crime was committed, the number of crimes committed, and the harm caused would be factors in determining the degree or severity of the commanders' failure and the punishment to be imposed, being evidence of the extent of their knowledge, negligence or recklessness and their culpability. As international criminal law is concerned with the most serious crimes of concern to the international community, the international courts should intervene at the top end of this graduated approach for the most serious 106 See Robinson, supra note 10, at 11, on how commanders are culpable when they negligently fail to fulfil their heightened duty to be vigilant and to supervise and be informed of their subordinates' actions. 107 Mettraux, supra note 12, at 53-4; Triffterer and Arnold, supra note 11, at 1087. While the basis of command responsibility is linked to the responsible command principle, which in turn is linked to international humanitarian law, command responsibility has been extended to other situations, for example, crimes against humanity and genocide outside of armed conflict and civilian leaders. Its potential application is much wider and arguably even extends to security firms and their managers, paramilitary leaders, and terrorist leaders. See Bantekas, supra note 72, at 1203-5; Mettraux, supra note 12, at 98-122. Similar reasoning could be potentially employed to provide a legal basis for extending command responsibility beyond the responsible command principle and international humanitarian law if the doctrine were classed as a separate offence.
derelictions of duty. Lesser derelictions could be addressed by the national system through criminal, civil or disciplinary means depending on their severity, nature and the harm caused.
Convicting commanders for their own failures would not only provide a stronger deontological argument that their punishment is 'deserved', but it would also satisfy the utilitarian arguments advanced to justify the extension of culpability under the doctrine. Commanders do not have to be convicted of the same crimes as their subordinates to achieve these objectives. If the commanders' dereliction of duty was a separate offence, they would still be held to account for failing in the responsibilities they assumed. They would be required to properly control their subordinates, to prevent crimes, and to enforce international humanitarian law. Convictions for dereliction of duty would deter commanders from failing to fulfil their duty and show subordinates that they cannot expect impunity. Significant penalties for gross failures would provide a similar incentive to commanders to monitor their subordinates' conduct; especially as the ex post assessment is still likely to be carried out by a judge with little or no military or combat experience.
The difficulties in proving the high-ranking commander's knowledge of and involvement in the crimes of their lower-ranking subordinate would also be alleviated. It would resolve the practical challenges of proof caused by 'hierarchical distance'. 108 Focusing on whether the commander fulfilled their duty, and not on the remote actions of another, would mean that the requisite evidence would be more readily apparent through their own action or omission. They would not 'slip between the crack' and escape liability. Indeed, it could expand the scope of liability as the type of commander or the relevant failure would not have to be restricted in order to make the doctrine more compatible with the culpability principle. Alignment of the doctrine's elements to the actual conviction would also resolve conundrums persisting under the present classification, such as the culpability of subsequent commanders who fail to punish crimes committed under the supervision of the previous commander. 109 However, reclassifying the doctrine would require the law to be revised, which is no easy feat. There may be substantial difficulties with respect to, for example, the lack of political will, challenges of enacting legal reforms, and the prohibition on the retrospective application of the law. These are significant concerns. Indeed, the difficulties have led some to maintain that reclassifying the doctrine is only a theoretical solution, as it is unlikely to be implemented in reality. 110 Yet, there was once considerable doubt about whether creating a permanent international criminal court was realistic and these political, legal, and pragmatic hurdles were overcome. 111 In addition, the alternative, as Bemba demonstrates, may be continued confusion and division. By contrast, reclassifying the doctrine as a separate offence could resolve the divergence between the crime and the conviction while heeding to the doctrine's justifications. It would uphold the culpability principle and afford an adequately wide scope to the doctrine, thereby preserving it as a valuable anti-impunity tool without compromising the high standards of liberal criminal justice.

Conclusion
The Bemba case demonstrates that despite decades of jurisprudence and academic critique, there is a lack of clarity and consistency regarding the nature of command responsibility and its core elements. Within the Bemba case alone, there are vastly different interpretations stretching from negligence to wilful failure and from there being no causation requirement to the requirement of 108 For example, Judges Monageng and Hofmański Opinion in Bemba allude to the practical difficulties of proving that high-ranking commanders had personal knowledge of all the specific details of the crimes of their low-ranking subordinates. See Judges Monageng and Hofmański Opinion, supra note 7, § 194. 109 See, e.g., Sander, supra note 17, at 112-13 (discussing the responsibility of subsequent commander). 110 See Ohlin, supra note 71, at 112-13. 111 See, e.g., P. D. Marquardt, 'Law Without Borders: The Constitutionality of an International Criminal Court', (1995) 33 Columbia Journal of Transnational Law 73. high probability that the omission contributed to the crime. This leads to significantly different standards and corresponding liability.
A key contributing factor to this division is disagreement on the very nature of the doctrine, namely whether command responsibility is a mode of liability or a separate offence of dereliction of duty. While international jurisprudence, including in Bemba, favours classing the doctrine as a mode of liability, interpretations under this classification can undermine the core culpability principle to varying degrees. At its more extreme end, the culpability principle could be severely undermined thus eroding a central tenet of criminal justicein the name of ending impunity for the most serious crimes. Other interpretations, including those which class it as a mode of liability, do not undermine the culpability principle or undermine it to a lesser extent. While more aligned with the objective of upholding high criminal justice standards, such interpretations can notably restrict the doctrine's scope and inhibit the effective prosecution of high-ranking commanders.
My objective was not to reconcile the competing interpretations of command responsibility or to argue that the separate offence approach represents an accurate interpretation of command responsibility as it currently stands. Rather, the purpose is to highlight that confusion regarding the doctrine persist and that reclassifying it as a separate offence could present a coherent solution. As a separate offence, there would be greater consistency between the doctrine's elements and the crime that the commander is convicted of, which could ensure that the culpability principle is observed. It would also provide command responsibility with an adequate scope that does justice to the international courts' objectives of combating impunity for serious crimes while upholding high standards of liberal criminal justice. The proposed approach would require a revision of the current law which is a monumental task. Yet, to continue down the present path runs the risk of continued disagreement, contradictions, and convolutions attended by the weakening of a core principle of criminal justice and the moral authority of international criminal law.