Melbourne Journal of International Law
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Melbourne Journal of International Law |
|
HOW COMMAND RESPONSIBILITY GOT SO COMPLICATED:
A CULPABILITY CONTRADICTION, ITS OBFUSCATION, AND A SIMPLE SOLUTION
DARRYL ROBINSON[*]
The literature on command responsibility is extensive and is rapidly growing more complex. In this article, I argue that command responsibility can be much simpler than it seems. I focus on a single puzzle, one hidden in plain sight. The puzzle is that international criminal tribunal jurisprudence uses command responsibility to convict persons without causal contribution to the crime, while also recognising a culpability principle that requires causal contribution. This stark contradiction has been obscured by many arguments in the jurisprudence and discourse. Indeed, many readers will raise a host of arguments to deny the contradiction I just described. I will dissect the major arguments to demonstrate that the contradiction does indeed exist. I argue that Tribunal jurisprudence took an early wrong turn in concluding that the ‘failure to punish’ branch of command responsibility is irreconcilable with a contribution requirement. This led to a rejection of causal contribution. Subsequent efforts to deny the resulting contradiction with the culpability principle, or to avoid it, have spawned many inconsistent, complex and convoluted claims about command responsibility. These include the descriptions of command responsibility as ‘sui generis’, as hybrid, as variegated, as responsibility for-the-acts-but-not-for-the-acts, as
neither-mode-nor-offence or as sometimes-mode-sometimes-offence. However, if we revisit the first misstep, a simple and elegant solution is available. Command responsibility is a mode of accessory liability; it requires causal contribution and it is perfectly workable. I draw on scholarship from criminal law theory to explore the parameters of the contribution requirement.
In this article, I will explore a particular puzzle, one hidden in plain sight. The puzzle is that Tribunal jurisprudence:[1] (1) recognises the principle of personal culpability, pursuant to which a person must contribute to a crime to be party to it; and (2) uses command responsibility to declare persons party to international crimes without a causal contribution. This is an internal contradiction. Many factors have obscured this basic contradiction. Indeed, many readers will promptly protest against the claim I just made, arguing for example that command responsibility is not a mode of liability, or that it is a new kind of mode of liability exempt from the contribution requirement. I will examine each of the major counter-arguments to demonstrate that the contradiction does indeed exist. The inquiry will generate insights about reasoning, about doctrine and about theory.
With respect to reasoning, I argue that Tribunal jurisprudence took an early wrong turn in its facile conclusion that the ‘failure to punish’ branch of command responsibility is irreconcilable with a contribution requirement. This led to the conclusion that contribution is not required. I will show that the early jurisprudence used a formalistic and doctrinal (source-driven) approach that did not engage with fundamental principles. The arguments against a contribution requirement did not even purport to address the contradiction with the system’s declared fundamental principles. The examination illustrates the need for more meaningful engagement with the principled constraints of justice. It is not enough to advance a utilitarian argument that a particular interpretation would make a doctrine more effective. We must also grapple with whether it is fair.
In a later phase of the discourse, as concern with liberal principles increased, more effort went into explaining away or resolving the contradiction between a mode that does not require causal contribution and a principle that modes require causal contribution. This generated increasingly convoluted claims about the nature of command responsibility. Tribunal-driven discourse has produced claims that command responsibility must not be a mode of liability, or that it is a new ‘sui generis’ form of liability whose nature has not yet been explained, or that command responsibility is sometimes a mode and sometimes a separate offence. It has led to descriptions of command responsibility that are complex, convoluted, contradictory, and almost mystical in their vagueness, a responsibility for-the-act-but-not-for-the-act, in an ill-defined twilight world that is neither mode nor offence. Such descriptions are elusively vague out of necessity, because clarity would reveal the contradiction.
With respect to doctrine, I argue that if we undo the first misstep, we immediately find a simple and elegant solution. Command responsibility in international criminal law (‘ICL’) is a mode of accessory liability and requires causal contribution. The solution instantly reconciles the Rome Statute of the International Criminal Court (‘Rome Statute’),[2] early case law and fundamental principles of criminal justice. No vague claims or complex regimes are required. Given that the Rome Statute expressly requires causal contribution, it already bypasses the problem that led to the complex command responsibility discourse in the first place. Thus, the International Criminal Court (‘ICC’) need not import most of that discourse.
With respect to theory, I provide a preliminary exploration of the theory of culpability and how it may apply in command responsibility. I draw on scholarship from criminal law theory to explore the parameters of the contribution requirement. It seems, for example, that relatively elastic ‘risk aggravation’ theories can suffice to satisfy the deontological commitment. I also explore possible deontologically-grounded arguments for culpability without causal contribution.
The syllogism which is at the core of my argument, and which is developed in Part III, is basically as follows:
1 ICL claims to comply with the fundamental principles of a liberal system of criminal justice.2 Those fundamental principles include the principle of personal culpability.
3 The principle of personal culpability requires that persons only be held liable for crimes to which they contributed.
4 Under the doctrine of command responsibility, the Tribunals and ICC hold the commander liable as a party (accessory) to the crimes of the subordinates, and charge, convict and sentence the commander as such.[3]
5 Therefore, to comply with our stated principles, command responsibility as a mode of liability must require that a commander’s dereliction contributed to the crimes of subordinates.
This syllogism is quite straightforward and proves a contradiction. However, that contradiction has been thoroughly obscured by numerous arguments and ambiguities in the jurisprudence. I will explore in turn each of the
counter-arguments that can be and have been advanced to resist this syllogism, in order to expose the problem more clearly.
First, one might regard principles such as the culpability principle as technical obstacles and irritants which hinder the aims of ICL, and thus seek to minimise or evade them using the same doctrinal techniques one uses with any inconvenient rule. In Part IV, I will argue that fundamental principles are rooted in our deontological commitment to respect individuals as moral agents rather than as mere objects to be acted upon for instrumental aims. This commitment is necessary for our endeavour to be described as ‘justice’.
Second, one might regard the doctrinal arguments in Tribunal jurisprudence as resolving the question. In Part V, I will show that the doctrinal arguments are internally unsound even on their own premises. More fundamentally, these doctrinal arguments are the wrong type of arguments, as they do not even attempt to answer the fundamental concern about culpability and the limits of principled punishment.
Third, a more sophisticated argument is that command responsibility is not a mode of participation in the underlying offence, but rather a distinct offence of its own. This is discussed in Part VI. The argument is an advance on the previous arguments, because it engages with fundamental principles. The approach has much to commend it, and I agree that criminalising command derelictions as a distinct offence would resolve the culpability concern, because the commander would no longer be held liable as a party to crimes to which she in no way contributed. However, I do not believe the ‘separate offence’ solution is available to the Tribunals or the ICC. Their applicable law seems to clearly establish command responsibility as a mode of liability. For judges to recognise a new separate crime would therefore raise serious concerns with the legality principle. Furthermore, the actual practice of the Tribunals unmistakably charges, convicts and sentences the commanders as parties to the underlying offences committed by subordinates. If the Tribunals are to hold the commander responsible as party to the core crimes, then they must contend with the culpability principle. Part VI also explores the related argument that command responsibility is a ‘sui generis’ mode of liability exempt from the requirement of causal contribution.
The fourth, and most ambitious, response would be to develop a new deontologically-grounded conception of culpability, in which causal contribution is not required. This possibility is briefly explored in Part VII. I conclude that arguments based on ‘accessory after the fact,’ subsequent ‘acknowledgement and adoption’, or ‘moral pollution’ do not provide convincing models.[4] There are however some intriguing proposals for a ‘ratification’ theory of culpability, whereby a person may ex post facto absorb liability for a deed by extending her will to it and ratifying it.[5] It is possible that a plausible account may be developed. However, some work remains to be done for ratification theory to support non-contributory derelictions as a mode of liability, especially in cases without subjective awareness of the crimes.
Finally, Part VIII explores the implications of the foregoing. The rejection of causal contribution in Tribunal jurisprudence is revealed to be an extraordinary proposition. Such a departure from culpability as generally understood and accepted by the system would require careful deontological justification, which Tribunal jurisprudence has not offered. Popular arguments to extend liability to successor commanders are even more problematic and in need of deontological justification. The restrained position of the International Criminal Tribunal for the Former Yugoslavia (‘ICTY’) Appeals Chamber majority in Prosecutor v Hadžihasanović (‘Hadžihasanović’)[6] has been much criticised, but should instead be commended as helping to reduce the culpability gap.
Conversely, the ICC is on a simpler and better path, because the Rome Statute expressly requires causal contribution. That provision has been criticised as an unfortunate and unnecessary limitation, but may be seen more charitably as keeping faith with principles rooted in treating persons fairly as moral agents. The Rome Statute avoids the problems that necessitated the contorted discourse over command responsibility, and thus the ICC should be careful about importing unnecessary aspects of that discourse. The first judicial treatment of the Rome Statute provision, in Prosecutor v Bemba (‘Bemba’),[7] adopts a ‘risk aggravation’ approach, which appears to be a plausible interpretation consistent with deontological thinking about culpability.
This article brings together ICL scholarship and criminal law theory literature on fundamental principles. For reasons of space it is impossible to provide a complete treatment of both bodies of work. The aim is simply to demonstrate convincingly that there is indeed a problem that warrants reflection and resolution. The article does not delve into the vast literature and jurisprudence on command responsibility, except insofar as is strictly needed to illuminate the causation issue. As for the scholarship on deontological principles, for reasons of space the article focuses on key works from the already vast literature in the English language. I caution that this language selection entails some emphasis on thinkers familiar with common law systems. Moreover, theoretical works in other languages may feature different debates, with different terms, identifying different problems and solutions. Accordingly, this work is simply an initial foray to illustrate a problem, and is the start of a broader conversation.
This article is unavoidably lengthy. I have striven to make it as short as possible. There are, unfortunately, a great many interconnected threads that must be untangled in order to unveil and explore the contradiction. Each issue is addressed as succinctly as can be done without losing clarity and substantiation. Different readers will have different preoccupations and objections that they will wish to see addressed first. I deal with the issues in a sequence that will provide the greatest clarity for the greatest number and will minimise repetition (repetition cannot be entirely eliminated given the intricate interconnection of the issues). I will cover each of the major objections — including doctrinal arguments, or the view that command responsibility is not a mode, or that it is a new kind of mode and hence is exempt from the contribution requirement, or that culpability can be reconceptualised — in the relevant section. When all is untangled, I hope to persuade you that there is a contradiction, that it has been obscured by the discourse, and that it has an elegant solution.
In this article, the term ‘command responsibility’ will be used, and reference will be made to military settings, but the broader concept of superior responsibility, which also encompasses civilian leaders, is acknowledged. The terms ‘core crime’ or ‘international crime’ refer to genocide, crimes against humanity and war crimes. The term ‘Tribunals’ refers to the international criminal tribunals for the former Yugoslavia and for Rwanda created by the Security Council.[8] Emphasis will be placed on the ICTY as it is the source of most of the relevant jurisprudence. The term ‘desert-based’ refers to justifications rooted in desert (ie, what the person can be said to ‘deserve’). Finally, as a counterbalance to the widespread use of the masculine pronoun, this article will use the feminine pronoun, especially in relation to commanders.
I will be using the term ‘doctrine’ and ‘doctrinal’ in the Anglo-American sense, in which the terms refer to standard legal reasoning and are contrasted with ‘theory’ or ‘theoretical’ considerations. This usage may differ significantly from the way the terms are used in other traditions, in which ‘doctrine’ includes theoretical considerations.[9] As the term is used here, ‘doctrinal’ analysis includes familiar techniques of textual interpretation, teleological argument and contextual argument. It also includes the techniques of relying upon or distinguishing past cases and precedents. Doctrinal analysis is concerned with what the sources require or permit. I contrast it in this paper with deontological analysis, which is concerned with a different question of what is fair (or, more precisely, what is consistent with our principles of justice). A theme of this article is that Tribunal jurisprudence has often responded to questions of fundamental principles with doctrinal analyses of statutes and precedents, and in doing so has missed the question of whether it is in compliance with its fundamental commitments.
The command responsibility doctrine, as articulated in the Statutes and jurisprudence of the Tribunals, imposes liability where:
(i) there is a superior–subordinate relationship;
(ii) the superior knew or had reason to know that a subordinate was about to commit crimes or had done so; and
(iii) the superior failed to take the necessary and reasonable measures to prevent such acts or to punish the perpetrators thereof.[10]
The Rome Statute takes a very similar approach, with minor differences.[11] First, and most importantly for present purposes, the Rome Statute requires that the commander’s dereliction contributed to crimes: ie, that the crimes were ‘a result of his or her failure to exercise control properly over such forces’. Second, while it is not the focus of this article, it should be mentioned that the Rome Statute also handles the mental element differently, drawing a distinction between military commanders and civilian superiors. The mental element for military commanders (‘should have known’) is similar to, but arguably slightly different from, the ICTY ‘had reason to know’ standard. Civilian superiors are accorded a more generous mental element, requiring that they ‘consciously disregarded’ information about crimes.[12]
This article analyses command responsibility as a mode of accessory liability, since that is how it has generally been understood and applied over the history of ICL. Some recent ICTY decisions suggest that command responsibility may instead constitute an entirely separate offence. This alternative conception will be discussed in Part VI. I will argue that command responsibility must be analysed as a mode of accessory liability, given the applicable law and given the practice of charging and convicting commanders as party to the core crime.
In what way does command responsibility reach beyond other modes of liability, doing something that other modes do not, thereby warranting its separate existence? Obviously, command responsibility has specific features and limitations, including that it applies only in a superior–subordinate relationship and focuses on omissions. But other modes of liability can also be applied in superior–subordinate relationships and can capture omissions. By comparing command responsibility with other modes of liability, we can tease out the distinctive value it adds. First, if a commander actually orders or instigates the crime, then she is liable by virtue of other modes of liability (such as ordering, instigating or joint commission through another). Second, if the commander does not order or initiate the crimes, but knows of the crimes and contributes to them, then she may still be liable through ‘aiding and abetting’ or other complicity doctrines.[13] Third, where the commander knows of the pending or ongoing crimes but nonetheless omits to prevent them, she can still be found complicit: for example, aiding and abetting by omission has been recognised where the person is under a duty to prevent crimes and in a position to act yet fails to do so.[14]
Accordingly, the distinctive reach of command responsibility is that it captures the commander who ‘had reason to know’ or ‘should have known’ of the crimes and failed to prevent or punish them.[15] Other modes of liability in ICL, such as aiding and abetting by omission, require knowledge of crimes; the modified mental element of command responsibility is the aspect that gives it additional reach. Command responsibility signals that, given the seriousness of the commander’s duties and the dangerousness of the activity of supervising troops, a deliberate or criminally negligent failure to fulfil the duty to control troops can be a basis for accessory liability in crimes resulting from that failure.
Tribunal jurisprudence claims that there is an additional difference: that the commander’s dereliction need not have any causal impact on the subordinate’s crimes. It is this claim that this paper will examine in detail.
As interest in fundamental principles has increased in the ICL literature, many features of command responsibility have come under criticism. Concerns have been raised about liability based on omission, the lack of requirement of actual knowledge, and the lack of requirement of causal contribution, giving rise to comparisons with ‘vicarious liability’ and concerns about imposing a stigma disproportionate to the moral wrongdoing.[16]
It is possible to defend command responsibility from many of these critiques. For example, liability based on omission is widely accepted in national legal systems, in theoretical literature and in ICL, provided that the person is under a legal duty to act, which is clearly the case for the commander.[17]
The concern about vicarious liability is more subtle. Vicarious liability is liability arising purely by virtue of a relationship, in the absence of the physical and mental requirements for complicity. Command responsibility can be distinguished from vicarious liability, because it does not flow merely from the fact of the relationship and the occurrence of crimes. Command responsibility requires personal fault on the part of the commander; namely, it requires that the commander knew or had reason to know of the crimes (mental element) and that the commander breached her duty to take reasonable steps to prevent and punish crimes (conduct element).[18]
Another concern raised in the literature is the modified mental element (‘had reason to know’ in the Statute of the International Criminal Tribunal for the Former Yugoslavia (‘ICTY Statute’)[19] and ‘should have known’ in the Rome Statute). We need not delve into the literature and jurisprudence on the precise meaning of the terms;[20] what is significant for present purposes is that the standard departs from the subjective mens rea standard commonly required in complicity. Complicity generally requires some level of subjective awareness of the criminal activity.[21]
While the mental element is not the focus of this article, a very brief outline of its possible justification is necessary, in order to anticipate some arguments below and to affirm the deontological viability of command responsibility. A useful framework for a desert-based account is provided by criminal law theorist Paul Robinson. In his study of ‘Imputed Criminal Liability’,[22] Robinson explored justificatory theories for a diverse class of inculpatory doctrines that hold persons liable even when they have not satisfied the ‘paradigm of liability’ for the offence (ie, satisfaction of all physical and mental elements).[23] One desert-based justificatory theory is ‘cumulative culpability’, where a person has caused the absence of a particular element (for example, caused the absence of knowledge) and has done so in a manner that is roughly ‘equivalently blameworthy’ to the paradigm of liability.[24]
The modified mental element for command responsibility might be justifiable under theories of equivalency or cumulative culpability.[25] Impressive steps in precisely this direction have been advanced by Jenny Martinez.[26] All other things being equal, negligence is of course generally considered less blameworthy than subjective mental states (such as intention and knowledge). However, intensified legal obligations are commonly placed upon persons who engage in inherently dangerous activities. The military commander is entrusted with the inherently dangerous activity of supervising persons with training in violence who have access to weapons and other equipment to carry out violence, and who have undergone indoctrination to reduce their inhibitions against violence. The law grants the commander privileges,[27] but it also requires her to be vigilant in remaining informed and taking measures to prevent and repress violations. Thus, the commander entrusted with such an inherently dangerous activity cannot argue that she was ‘merely’ criminally negligent in creating her own ignorance. Her indifference, in the context of her responsible relation to a clear public danger, is, arguably, sufficiently blameworthy in a desert-based account.[28]
If the foregoing is correct, then the mental element of command responsibility may differ from complicity doctrines known in some national laws, but it is not a departure from the underlying deontological commitment to treat persons in accordance with desert.[29] The modified mental element is rooted in individual desert, in light of the role and responsibilities assumed by the commander and the dangerousness of the activity. The concept of complicity by omission (by those under a duty to prevent crimes) is already established; command responsibility affirms that, given the control over danger, a criminally negligent failure to know of subordinate crimes is a sufficiently blameworthy state of mind to ground accessory liability.[30]
The focus of this article is the question of causal contribution; namely, whether the dereliction(s) of the commander must contribute to or facilitate the subordinates’ crimes in order for her to be an accessory to those crimes. Tribunal jurisprudence on command responsibility rejects causal contribution, whereas the Rome Statute embraces the requirement. This section will show that Tribunal jurisprudence is problematic as it: embraces the culpability principle; acknowledges that this entails that a person must contribute to a crime to be liable for it; and yet allows conviction of a person as a party to a core crime by virtue of her failure to punish others, even if her failure did not facilitate or encourage any core crime. We will explore some scenarios where this is problematic: the isolated crime, the initial crime, and the successor commander.
One of the fundamental principles of a liberal system is the principle of personal culpability. The principle of personal culpability has both objective aspects (a personal connection to the crime) and subjective aspects (a blameworthy mental state). Our focus in this paper is the objective aspect: ie, that we hold persons responsible only for their own conduct and the consequences thereof.[31] Culpability is personal, hence we cannot punish a person for the crimes of others in which she was not involved.
Obviously, criminality often involves multiple actors, each contributing to a crime in different ways and in differing degrees. Thus, an individual may share liability relating to acts physically perpetrated by others, provided that the individual contributed to the acts and did so with the requisite mental state.[32] Punishing persons only for their own wrongdoing means that the accused must contribute in some way to a crime to be liable for it. ICL scholars Guénaël Mettraux and Ilias Bantekas have observed, respectively, that the requirement that the accused be ‘causally linked to the crime itself is a general and fundamental requirement of criminal law’[33] and that ‘in all criminal justice systems, some form of causality is required’.[34]
Those parties to a crime who are most directly responsible are liable as principals, and other more indirect contributors are liable as accessories. Accessory liability is a well-established means of derivative liability.[35] Of course, not every legal system expressly distinguishes between principals and accessories as a matter of doctrine, but we refer here to the distinction as a conceptual category, one which is recognised in ICL.[36] Principals make an essential contribution, often expressed as sine qua non or ‘but for’ causation of some aspect of the crime.[37] For example, an archetypal principal is the physical perpetrator, who brings about the actus reus directly through her own voluntary acts. Other types of principal make an essential contribution in different ways.[38] By contrast, the contribution of an accessory may be more indirect: the accessory’s actions either influence or assist the voluntary acts and choices of the principal(s).[39] Thus, principals cause the crime, whereas accessories influence or assist the principals.[40]
The ‘contribution’ requirement for the accomplice is more elastic than a strict (‘but for’) ‘causation’ test for several reasons. One reason, as will be discussed further in Part VII, is that accomplices are liable by virtue of their influence on perpetrators, and there are special considerations in tracing ‘causation’ or even ‘causal contribution’ through the voluntary and informed acts of others. Accordingly, it is not required that an accessory ‘cause’ the crime in the sense of a sine qua non or ‘but for’ causal relation; all that is required is some ‘contribution’. As Michael Moore writes, ‘[t]o be an accomplice, my act must have something to do with why, how or with what ease the legally prohibited result was brought about by someone else’.[41] Part VII will engage in closer principled analysis of this question, and we will see that there are different plausible formulations of the requisite degree of contribution, including that it facilitated or had an effect on the crime, that it at least ‘could have’ made a difference, or that it increased the risk of the crime occurring.[42] For now, the significant point is that Tribunal jurisprudence on command responsibility, particularly with respect to a commander’s ‘failure to punish’, fails to satisfy any formulation of the requirement.
ICL is conceived and presented as a liberal system of criminal justice, and indeed one that prides itself as a ‘model’ respecting the fundamental principles of a liberal justice system in an exemplary manner.[43] Accordingly, ICL states its compliance with the principle of culpability. For example, in Prosecutor v Tadić[44] it was recognised that
the foundation of criminal responsibility is the principle of personal culpability: nobody may be held criminally responsible for acts or transactions in which he has not personally engaged or in some other way participated (nulla poena sine culpa).[45]
In conformity with this principle, ICL jurisprudence recognises the general proposition that accessory liability requires some contribution to the underlying crime. For example, the International Criminal Tribunal for Rwanda (‘ICTR’) in Prosecutor v Kayishema affirmed that it is ‘firmly established that for the accused to be criminally culpable his conduct must … have contributed to, or have had an effect on, the commission of the crime’.[46] Similarly, ICTY jurisprudence has confirmed that ‘rendering a substantial contribution to the commission of a crime is indeed expressing a feature which is common to all forms of participation’.[47] Tribunal jurisprudence has also recognised that conduct after the completion of a crime cannot be regarded as contributing to the commission of the crime.[48]
A typical, and plausible, elaboration on the contribution requirement is provided in Prosecutor v Orić (‘Orić’).[49] The decision held that it is ‘enough to make the performance of the crime possible or at least easier’[50] and that the contribution can be any assistance or support, whether present or removed in place and time, furthering or facilitating the performance of the crime, provided that it is given ‘prior to the full completion of the crime’.[51] The decision confirmed that the contribution may be in the form of an omission, if the accused was under an obligation to prevent the crime.[52] The Orić decision also confirmed that the test is not a ‘but for’ test, but merely that there be a substantial or significant effect which furthers or facilitates the commission of the crime.[53]
After affirming the culpability principle and the contribution requirement entailed therein, Tribunal jurisprudence nonetheless goes on to assert that the requirement does not apply to command responsibility. For example, the Orić decision acknowledges that modes of liability require a causal contribution and thus that superior responsibility would ‘require a causal contribution to the principal crime’, yet asserts that causal contribution is not required ‘for good reasons’.[54] The quality of those reasons will be scrutinised below in Part V. In the following sections I will introduce the emergence of the no-contribution position and its implications.
The doctrine of command responsibility provides two distinct ways to prove the dereliction by the commander: ‘failure to prevent’ and ‘failure to punish’.[55] The first branch requires that the commander ‘failed to take the necessary and reasonable measures’ to prevent the crimes.[56] The first branch generally does not pose a significant culpability problem. Where the commander failed to take reasonable steps to prevent the crime, this implies at least a potential contribution or risk aggravation, which satisfies plausible formulations of the contribution requirement, as will be discussed in Part VII. Given that the commander has a duty to provide training and establish preventive systems, the failure to do so can be seen to facilitate crimes in comparison with the situation that would exist had the duty been met.[57]
It is the second branch, the ‘failure to punish’, that has caused confusion and difficulty. This branch refers to the failure of the commander to take the reasonable and necessary measures to investigate and punish or to refer the matter to competent authorities for investigation and prosecution.[58] A commander’s failure to punish for a particular crime obviously can only occur after that crime, and hence cannot causally contribute to that crime. For this reason, Tribunal jurisprudence has declared that it is ‘illogical’[59] and ‘would make no sense’[60] to require that the failure to punish the crime contribute to that same crime. From this observation, the Tribunal reasoned that ‘the very existence’ of the failure to punish branch in art 7(3) ‘demonstrates the absence of a requirement of causality’.[61] Accordingly, the ICTY rejected the contribution requirement.[62]
While the first step in the Tribunal’s chain of reasoning was impeccable, the second step was not. The Tribunal is indeed correct that a failure to punish a crime cannot retroactively causally contribute to that same crime. However, this does not demonstrate that the ‘failure to punish’ branch is incompatible with the contribution requirement. The argument only seems viable if we fail to contemplate the possibility of a series of crimes.
Consider the scenario where subordinates commit not one crime but a series of crimes, which is indeed the typical situation in ICL. The first crime or crimes are committed. At some point the commander either learns of the crimes or has enough information to be wilfully blind or criminally negligent in failing to know.[63] The commander fails to take reasonable steps to have the crimes investigated and prosecuted, and crimes continue to occur. Although this failure of the commander cannot retroactively contribute to the initial crimes, it can and does contribute to each subsequent crime. Her failure to punish the prior crimes facilitates the subsequent crimes, especially in comparison to the legally expected baseline of her vigilant investigation and prosecution of crimes. Her dereliction facilitates and elevates the risk of crimes in two ways: in terms of rational calculations, because subordinates perceive a reduced risk of punishment, and in terms of expressive function, because she fails to instil the appropriate values by repudiating the crime. In such cases, the commander can properly share in accessory liability for the subsequent crimes, because her failure to punish past crimes is a culpable omission which contributed to subsequent crimes.
By considering the (common) scenario of multiple crimes, we see that the ‘failure to punish’ branch can perfectly well be reconciled with a requirement of causal contribution. Hence, there was no incompatibility or contradiction that would require, or even permit, the Tribunal to dispense with a basic requirement of the culpability principle. With this faulty reasoning, the Tribunal abandoned the contribution requirement, which is required by the culpability principle, unnecessarily and all too lightly. Had the Tribunals approached the provision with the culpability principle in mind, the provision could readily have been interpreted compatibly with the requirement. Indeed, the Rome Statute, by contrast, specifically recognises and respects the contribution requirement.
Many readers will object to my assertion that the Tribunal abandoned the culpability principle. The most significant counter-arguments will be that causal contribution is not legally required, or that command responsibility is a separate offence, or that it is a special new mode exempt from the contribution requirement, or that culpability must be reconceived. I will address each of these arguments below.
One could object that requiring contribution sounds cumbersome because a conviction based on failure to punish is only feasible if that failure contributed to subsequent crimes.[64] One might feel that it is a weakness that we cannot hold the commander liable for the initial crimes or for an isolated crime by virtue of her failure to punish.[65] However, the problem remains that she did not facilitate or have an effect on those crimes.
If the following premises are correct:
(i) ICL is a liberal system;
(ii) a liberal system entails the principle of culpability;
(iii) the principle of culpability requires a causal contribution;
then one has no choice but to grapple with this requirement. Accordingly, one would have to either:
(a) accept the requirement of causal contribution (as the Rome Statute does);
(b) develop a convincing deontological justification for the apparent departure;[66] or else,
(c) acknowledge that Tribunal jurisprudence contravenes a fundamental principle and thus forego claims of exemplary compliance with such principles.
To illuminate the implications of allowing convictions without causal contributions, I will outline two scenarios of non-contributory failures to punish. One is the problem of the isolated crime and the other is the problem of the successor commander.
The first problem arises where a commander is in charge of troops and a crime occurs, but the crime was not facilitated, encouraged or affected by any failure of the commander to prevent or punish. This scenario can arise only where the commander has adequately satisfied her ‘preventive’ duties. If she breached her duty to prevent, then the contribution requirement would be met for the single crime and she could be held liable in relation to that crime.[67] Accordingly, the scenario of the non-contributory dereliction necessarily entails that the commander has adequately satisfied her duties under the ‘prevention’ branch, by arranging appropriate training, appropriate orders and a system of supervision.
Nonetheless, a subordinate commits a war crime. The commander knows or has reason to know of the crime, but fails to investigate, punish or refer the matter to competent authorities. No further crimes occur.[68] The commander has clearly failed in her responsibilities, and she may face various consequences for her dereliction, but we cannot convict her as a party to the core crime. She has not contributed to or had an effect on the core crime. The culpability principle, which ICL claims to respect, requires causal contribution and thus precludes accessory liability for the subordinate’s crime.
A common objection, if one adopts a purely utilitarian perspective of maximising constraints on commanders in order to maximise victim protection, is that the constraint of requiring contribution would create a ‘gap’ that will allow commanders to ‘escape justice’ in such a scenario.[69] However, such arguments beg the important question of whether conviction in such circumstances would constitute ‘justice’. Justice requires us to consider what the accused deserves, which brings us unavoidably to the question of culpability. If the commander is not culpable for the core crime, then it is our inability to convict for that crime which constitutes ‘justice’. Furthermore, there is also a separate but serious question of mandate. Given that the mandate of the Tribunals is to deal with the persons most responsible for the most serious international crimes, it is questionable whether they should be occupied with derelictions that did not make and could not have made any contribution to international crimes.[70]
Another common objection to the contribution requirement is that the scope of criminal liability would fail to reflect the full scope of the humanitarian law duty, which requires the commander to punish past crimes regardless of whether she contributed to them.[71] Indeed, the commander clearly does have a duty under humanitarian law to punish past crimes, regardless of whether she contributed to them.[72] Her failure to punish would breach humanitarian law. However, this does not demonstrate that we can hold her personally criminally liable as an accessory to those crimes. I have discussed elsewhere the fallacy of assuming
the co-extensiveness of criminal law norms with human rights or humanitarian law norms, without reflecting on fundamental principles applicable in a criminal law context.[73] The commander’s failure to punish may have ramifications under humanitarian law for her state or armed group. She may also personally face criminal law repercussions, if a lawmaker with jurisdiction has criminalised
non-contributory derelictions of duty.[74] In such a case, she would be prosecuted for what she did (for example, her dereliction in failing to punish) and not for what the subordinates did (for example, genocide).
We cannot convict the commander as party to genocide by virtue of a failure to punish the crimes of others, if that failure did not contribute to any crimes of genocide. She neither participated in nor contributed to any crimes of genocide. She is not culpable for genocide. ‘Party to genocide’ is the wrong label to describe her misdeed to the world. In the absence of any contribution to the crimes, to convict a person for genocide, crimes against humanity or war crimes, and to impose the stigma that such crimes bear, contradicts the principle of culpability that ICL claims to respect.[75]
An even more glaring problem of non-contributory dereliction arises in the scenario of the ‘successor commander’. In Hadžihasanović, the ICTY considered the situation of a commander, Kubura, who had taken up his command position after certain crimes were committed. Kubura was nonetheless charged with crimes committed prior to his assignment, by virtue of command responsibility and his failure to punish those crimes once he took up the post.
The prosecution, the Trial Chamber and the two dissenting judges in the Appeals Chamber took the proposition that causal contribution is not required and pressed it to its furthest logical conclusion.[76] If no causal contribution is required, then it follows that the accused need not even have been in command or involved in the outfit at the time of the crimes — indeed, if we continue to follow the implications, the accused need not even have been born at the time of the crime — all that matters is that the accused at some point assumed command, became aware of past crimes or had reason to know of them, and failed to punish the persons responsible. Such a scenario would meet all the doctrinal requirements of art 7(3) of the ICTY Statute, if we apply them mechanistically and without any regard for fundamental principles.
On appeal, a bare 3:2 majority of the Appeals Chamber rejected successor commander liability, over some strong dissents and with some heated judicial language on all sides.[77] The majority held that the commander must at least have been in command at the time of the crimes. The reasoning of the majority was not explicitly based on concern for the culpability principle but rather on the more formalistic and doctrinal grounds that prior sources and authorities did not seem to support command responsibility convictions of successor commanders.[78] Judges Shahabuddeen and Hunt, in dissent, would have allowed successor commander liability.[79]
A large body of literature has developed on the distinct but related issues of successor commander liability and the nature of command responsibility.[80] Rather than receiving applause for its principled restraint, the majority position has come under considerable criticism. While some scholars side with the majority, many argue that the majority position fails to understand command responsibility and creates a ‘loophole’ and a ‘gaping hole’ through which ‘wrongdoers will escape reproach’.[81] Within the ICTY and the ICTR, Trial Chambers have openly expressed their discontent with the majority decision.[82]
A Trial Chamber of the Special Court for Sierra Leone declined to follow the majority approach and instead sided with the approach of the dissenting judges.[83] The ICTY Appeals Chamber itself almost overturned the majority position in the later decision of Orić.[84] Separate opinions in the Orić appeal decision described the Hadžihasanović majority decision as an ‘erroneous decision’, ‘highly questionable’ and an ‘arbitrary limitation’[85] and noted that there ‘is a new majority of appellate thought’.[86] The Appeals Chamber narrowly declined to overturn Hadžihasanović on the grounds that facts in the Orić appeal did not squarely require a determination on that issue.[87]
What appears to be largely missing from the judicial conversation is the observation that holding a person as party to crimes that were completed before she even joined the unit is quite a remarkable proposition. It appears to be a startling departure from the stated commitment of the system to the culpability principle. If the proposition is to be entertained at all, it would require a new understanding of culpability, backed by some meticulous deontological justification.
The culpability problem is not entirely overlooked. For example, Judge Shahabuddeen acknowledges that a mode of liability requires causal contribution.[88] His solution is to characterise command responsibility as a separate offence. Part VI will examine this characterisation. I will argue that the ‘separate offence’ approach would indeed avoid the culpability problem, but this solution does not seem to be available to the judges of the Tribunals and the ICC, in light of their applicable law. Moreover, Tribunal practice does in fact declare commanders guilty as party to the underlying crime, see Part VI.
The position taken by the prosecution[89] and apparently by most of the jurisprudence,[90] involves a starker contradiction, because it regards command responsibility as a mode of accessory liability, rejects the contribution requirement, and yet regards Tribunal law as compliant with the culpability principle. This position could only be rescued by some careful deontological justification of a new concept of culpability. Yet Tribunal jurisprudence seems genuinely unaware that there is a problem requiring such justification.
The culpability problem is not immediately evident, because several issues obscure the problem and the need for justification. These issues include ambivalence about the role and applicability of fundamental principles (Part IV), doctrinal arguments that the causation requirement simply does not apply in command responsibility (Part V), and ambiguity about the nature of command responsibility (Part VI). The following Parts will address these issues, to bring the problem into clearer relief.
A first argument against the culpability concern would be to question why we need to take the culpability principle seriously in the first place. Some may see fundamental principles as unfortunate doctrinal constraints that hinder the maximisation of justice for victims. Hence one could understandably seek to minimise or evade such principles using the types of arguments used to avoid inconvenient doctrines.
It is useful to recall why a liberal system of criminal justice respects fundamental principles. A liberal system embraces principled constraints because of its respect for the autonomy and personhood of the individuals subject to the system. As H L A Hart famously showed, even if the general justifying aim of the criminal law system as a whole may be utilitarian (protecting society), the question of ‘[w]hat justifies us in applying the system of punishment to a particular individual?’[91] cannot be decided solely on utilitarian grounds.[92] Utilitarian approaches do not provide internal constraints to prohibit, for example, punishing the innocent, if doing so were found to be net beneficial for social welfare. More fundamentally, consequentialist arguments simply fail to capture why we abhor imposing punishments that are not deserved: it is not the inefficiency; it is that it is unjust.[93]
A liberal system acknowledges human beings as moral agents, possessed of dignity and personhood and capable of directing their behaviour by reason. The conviction that the accused cannot be used solely as a ‘means’ to a societal ‘end’[94] is the core commitment of a liberal criminal justice system. Persons are entitled to respect for their autonomy except to the extent interference is justified. This basic commitment translates into certain fundamental principles, which include the principle of personal culpability. This is why it is not sufficient to provide utilitarian arguments showing that punishing individuals in certain circumstances would bring societal benefits (such as deterrence); we also require a deontological justification, showing that the punishment is ‘deserved’.
The issue is not merely one of doctrinal coherence but of philosophical coherence. ICL is a liberal project aimed at upholding human dignity and autonomy. If ICL wishes to instil the value that human beings must be treated as moral agents possessed of dignity, it must in turn treat persons as moral agents possessed of dignity, and respect the according constraints.[95] To treat persons as objects in order to advance a message that persons may not be used as objects would be to embark on a project riven with self-contradiction. While fundamental principles may at times seem to inhibit the pursuit of maximal victim protection, the alternative — to create a punitive system for the ‘administrative elimination of wrongdoers’[96] in the name of advancing human rights — seems philosophically incoherent.[97]
Some very compelling critiques of the liberal justice model have been advanced in the literature. For example, scholars such as Mark Drumbl and Mark Osiel have convincingly argued against the slavish replication of national doctrinal frameworks in ICL.[98] I agree that ICL need not replicate particular doctrines, or even particular articulations of fundamental principles, merely because they appear in national systems.[99] While agreeing with the argument, I would add the caveat that this does not mean that we are free to abandon the underlying deontological commitment to the human beings that may be subject to the ICL system. As I explain in more detail elsewhere,[100] the special contexts and special challenges of ICL may trigger new realisations about the nature of our deontological commitment, and thus properly lead to new articulations of fundamental principles.[101] Ultimately, however, we must still furnish deontological justifications for ICL doctrines.
A second way to overlook the contribution problem is to regard the doctrinal arguments in Tribunal jurisprudence as satisfactorily settling the question. The three main arguments advanced in Tribunal jurisprudence are:
(i) that precedents do not require causation;
(ii) that the ‘failure to punish’ branch is incompatible with a causation requirement; and
(iii) that requiring causation would render command responsibility superfluous in light of the other modes of liability.
I will argue that these arguments are questionable even as doctrinal arguments, in that their premises are internally unsound. More fundamentally, these doctrinal arguments are the wrong type of answer, in that they do not even attempt to answer the concern that the system is contradicting its recognised fundamental principles.[102]
I emphatically acknowledge that the Tribunals have operated in a pioneering phase of ICL, and confronted the massive and complex task of constructing doctrine from diverse authorities. In elaborating the rules, they faced countless legal questions that were substantial in scope. It is entirely understandable that they could not give detailed consideration to every fine point, especially in the earliest cases. It is largely thanks to the Tribunals that we now have a corpus of law, and are able to take a step back to critically assess the corpus. Our task is critically to assess the corpus and the reasoning employed, in order to improve upon them.
A main line of argument in the Tribunal’s rejection of a contribution requirement has been that past precedent did not require it. For example, in Prosecutor v Delalić (‘Čelebići’),[103] the defence argued that a failure to punish must causally contribute to the commission of subsequent criminal acts.[104] The Trial Chamber acknowledged ‘the central place assumed by the principle of causation in criminal law’,[105] but nonetheless held that a causal contribution ‘has not traditionally been postulated’ as a condition for liability under command responsibility.[106] In a one sentence analysis, the Chamber asserted that it ‘found no support’ for a requirement of causal contribution for command responsibility in the case law, treaty law or (with one exception) the literature.[107]
Similar defence arguments were rehearsed before the ICTY Appeals Chamber in Prosecutor v Blaškić (‘Blaškić’).[108] The Appeals Chamber cited with approval the analysis in Čelebići and rejected the argument that causal contribution is an element of command responsibility.[109] In a related section, the Appeals Chamber considered two post-World War II cases and found that they did not stipulate any requirement of causal contribution.[110] Subsequent cases have, in turn, cited Čelebići and Blaškić as authority settling this question.[111]
The problem with these references to past authority as a response is that they constitute the wrong type of answer. We may distinguish broadly between two tasks.[112] One task is the formalist doctrinal task of deriving rules through deduction and induction from the appropriate legal sources, using interpretive tools, teleological arguments, and contextual arguments to promote a coherent schema from the sources. The second task is analysis of compatibility with fundamental principles. The second task requires a deontological analysis, a consideration of the fundamental principles that limit our license to punish individuals. The task requires more than a formalistic interpolation of what the sources allow us to do; it requires an assessment of the fairness of those rules. To say that there is precedent or authority for a doctrine simply fails to answer the challenge that the doctrine contradicts a fundamental principle.[113]
In fairness to the precedent-based reasoning in the Čelebići and Blaškić decisions, it must be acknowledged that defence lawyers in those cases primarily characterised their challenge as one based on the principle of legality (nullum crimen sine lege). Hence, reference to doctrine was an appropriate response to address that challenge. The problem is that subsequent Chambers have regarded Čelebići and Blaškić as conclusively settling the issue, and hence have failed to engage with the culpability challenge. Reference to precedent is not an answer to the question of whether the doctrine complies with the principle of culpability, which is one of the avowed principles of ICL.
While this article will not embark upon the vast doctrinal question of whether past authorities did or did not support a requirement of causal contribution, it is worth noting that the question was much more open than the Chambers appeared to appreciate. Several scholars have shown that past cases and authorities provide considerable support for a contribution requirement.[114]
Indeed, the Čelebići decision itself cites passages from authorities that expressly indicate that, even under the ‘fail to punish’ branch, a causal contribution to crimes is required. To take just two examples, Čelebići cites the post-World War II decision, United States v Toyoda, which described the principle as covering the commander who ‘by his failure to take any action to punish the perpetrators, permitted the atrocities to continue’.[115] Similarly, Čelebići cites legislation of the former Yugoslavia which states that ‘[a] military commander is responsible as a participant or an instigator if, by not taking measures against subordinates who violate the law of war, he allows his subordinate units to continue to commit the acts’.[116] These and other authorities show an understanding on the part of other courts and lawmakers that liability arises where the commander’s failure to punish contributed to crimes by permitting them to continue; an understanding which is consistent with the principle of culpability.
Thus, it is puzzling that the Čelebići Trial Chamber nonetheless managed to find ‘no support’ for a contribution requirement. It is possible that if the Tribunal had approached the doctrinal question with more sensitivity to the fundamental principles — ie, looking at the precedents against the fundamental backdrop of personal culpability — they might have discerned more support for a causal contribution requirement in the authorities, including in the very passages that they cited.
As discussed above in Part III, the Tribunals have perceived an incompatibility between the existence of the ‘failure to punish’ branch in the Rome Statute text and a causation requirement. In Čelebići, the defence argued that a ‘failure to punish’ should give rise to accessory liability only if that failure is ‘the cause of future offences’.[117] The Trial Chamber appears to have missed the subtlety of the defence argument, and focused instead on whether a failure to punish a crime can cause that same crime. The Trial Chamber held that ‘no such causal link can possibly exist’ between a failure to punish an offence and ‘that same offence’.[118] Thus, ‘the very existence’ of the failure to punish branch in art 7(3) ‘demonstrates the absence of a requirement of causality’.[119]
The reasoning is sound as far as it goes, but it is too simplistic. The defence was not arguing for the proposition of retroactive causation. Rather, the defence argument was that a failure to punish can create accessory liability only with respect to subsequent crimes encouraged or facilitated by that failure.[120] The accused could then properly be held liable for the subsequent crimes because she contributed, with fault, to those crimes. Hence, there is nothing illogical about recognising the ‘failure to prevent’ branch while also respecting the contribution requirement.
It is also often argued that recognising the contribution requirement would render the ‘failure to punish’ branch redundant, because all breaches would be a failure to ‘prevent’.[121] However, ‘failure to prevent’ and ‘failure to punish’ are not redundant; they offer two distinct ways to prove the failure of the commander. A prosecutor may prove either inadequate preventative measures or inadequate efforts to investigate and prosecute crimes. Either provides the dereliction that, if accompanied by a blameworthy state of mind and a contribution to crimes, can ground accomplice liability for resulting crimes. Thus, respecting the contribution requirement does not render the ‘failure to punish’ branch superfluous.
The faulty ‘incompatibility’ argument was a major factor putting the Tribunal onto its path against contribution, which then spawned complex and contradictory assertions on how to reconcile the position with fundamental principles and increasingly complex and contradictory assertions about the nature of command responsibility. The path is now probably too well-trodden for the Tribunals to reverse course. However, other international and national courts should inspect the issue afresh; indeed, that one faulty argument has generated an increasingly complex and convoluted discourse that need not be replicated.
Another common argument against a contribution requirement is that it would render command responsibility redundant with other modes of liability. The decisions in Prosecutor v Halilović (‘Halilović’)[122] and Orić held that the absence of a contribution requirement is the distinguishing feature of command responsibility.[123] Thus, they argue, if causal contribution were required, then the ‘borderline between article 7(3) [command responsibility] and … 7(1) [the other modes] … would be transgressed and, thus, superior criminal responsibility would become superfluous’.[124] However, this argument overlooks that command responsibility is already quite distinct from other modes of liability by virtue of the modified mental element; it allows conviction based on a ‘had reason to know’ or ‘should have known’ standard.[125] Hence, it is not true that respecting the contribution requirement (and the culpability principle) would render command responsibility superfluous.
A related argument is that ‘[i]f a causal link were required this would change the basis of command responsibility’ because ‘it would practically require involvement on the part of the commander … thus altering the very nature of the liability imposed under Article 7(3)’.[126] However, upholding the contribution requirement does not require ‘involvement’ by the commander; the essence of command responsibility remains the failure to become involved where there was a duty to do so. The failure to intervene facilitates the crime in comparison with the situation that would have existed if the commander had met her duty (ie, if she had taken adequate measures to try to prevent and punish crimes).[127] This crime-facilitating effect of the commander’s failure satisfies the contribution requirement.
More fundamentally, these arguments are not only doctrinally unsound for the reasons just given, they are also unsound as a form of argument because they fail to address the objection. The objection is that causal contribution is required in order to respect the principle of culpability. ICL claims to respect the principle of personal culpability as ‘the foundation of criminal responsibility’ and thus to only hold persons responsible for transactions in which they ‘personally engaged or in some other way participated’.[128] Technical doctrinal arguments, such as reconciling one provision with another, are not an answer to the challenge that one is contradicting one’s stated fundamental principles. To answer such a challenge, one has to look up from the tools of textual construction and address whether the doctrine complies with the system’s concept of culpability.[129]
A third response to the culpability problem is to assert that command responsibility is not a mode of liability at all. This response is more sophisticated than the previous ones, because it does not ignore the culpability principle; it avoids the application of the principle by denying that command responsibility is a mode. In Hadžihasanović, the Appeals Chamber confronted the scenario of the successor commander, a scenario which placed the problems of not requiring causation in particularly stark relief. Faced with defence objections to liability in the absence of ‘any involvement whatsoever in the actus reus’,[130] Judge Shahabuddeen advanced an innovative solution:
I prefer to interpret the provision as making the commander guilty for failing in his supervisory capacity to take the necessary corrective action … Reading the provision reasonably, it could not have been designed to make the commander a party to the particular crime committed by his subordinate.[131]
This approach has been seized upon in several subsequent decisions,[132] and there is now a great controversy over the nature of command responsibility.[133]
I will advance two points. First, punishing derelictions as a separate offence would indeed solve the culpability problem, because the commander would be charged only for her own dereliction and not treated as a party to crimes to which she did not contribute. Second, while lawmakers in general are free to adopt that solution, this solution does not appear to be available to the judges of the Tribunals and ICC, given their Statutes and applicable law. The solution would avoid the culpability problem but at the price of an even more disconcerting legality problem. Moreover, actual Tribunal practice does indeed charge and convict persons for the underlying crimes and hence expresses to the world their liability for these crimes.
Before proceeding to analyse the ICTY approach, I must acknowledge an alternative and more sophisticated argument for a separate offence approach. Kai Ambos and others have argued for a separate offence interpretation, not as a device to justify convictions of successor commanders, but for the principled reason that it is the only way to comply with liberal principles.[134] This argument has commendable plausibility if there is no way to justify command responsibility as a mode of liability. If such an interpretation were the only way to conform to fundamental principles, then contextual and teleological interpretation might indeed support a strained textual reading in order to conform to fundamental values.[135] However, I would suggest that the ‘separate offence’ reading is not the only way to comply with fundamental principles and hence a counter-textual interpretation is not necessitated. Culpability concerns can be addressed by recognising the contribution requirement, as the Rome Statute expressly does. The departure from the normal subjective mental element may also raise concerns; I outlined briefly above how this concern might be addressed (see Part II(B)).
If breach of command responsibility were legally posited as an entirely separate offence, then the concerns raised here about culpability and fair labelling would evaporate immediately.[136] The culpability problem would be resolved, because the commander would not be held indirectly liable as a party to crimes to which she in no way contributed. Instead, she would be held directly liable for her own dereliction. There are two ways this crime could be defined. First, the dereliction itself can be conceived as the entirety of the offence and punished as such. On this approach, the definition of the crime need not require that the dereliction contributed to core crimes by others or even that core crimes occurred at all. Alternatively, one could regard breach of command responsibility as a separate offence and yet still require causation; thus the crime would be concerned with derelictions that contribute to core crimes by others.[137]
Therefore, a lawmaker can certainly recognise derelictions as offences in their own right, with or without a contribution to core crimes. In the latter case no causal contribution need be shown, as the dereliction itself would constitute the offence. Furthermore, in terms of fair labelling and the expressive function of criminal law, the commander would not be presented to the world as a party to ‘genocide’; instead she would be convicted with a label that accurately communicates her wrongdoing.
This paper does not enter into the debate as to whether the ‘mode of liability’ approach or the ‘dereliction’ approach is preferable. Indeed, a lawmaking authority could even legislate both concepts, recognising command responsibility as a mode of participation and also establishing a separate dereliction offence. The German legislation is a commendable model.[138] The ‘mode’ approach has the advantage of expressing the solemnity of the duty to control troops and reflecting the grave harm flowing from the commander’s failure to do so. The ‘dereliction’ approach has the advantage of capturing failures that do not meet the requirements for accessory liability in the principal’s crime.
While the ‘separate offence’ approach would indeed address the culpability problem, it is doubted that the option is available to the Tribunals. First, the applicable law does not seem to create the offence or give authority to establish it. Second, the actual charges and convictions entered by the Tribunals do in fact hold persons responsible as party to the crimes.
The texts of the Statutes appear to recognise command responsibility as a mode of liability and not as a crime. Structurally, the Statutes do not include command responsibility among the definitions of crimes, but include it instead among the ‘general principles’. For example, in the ICTY Statute, the crimes are listed in arts 2–5, whereas command responsibility appears in art 7, which contains principles of ‘individual criminal responsibility’, including the other modes of liability, such as planning, instigating, ordering and aiding and abetting.[139] Similarly, in the Rome Statute, definitions of crimes appear in pt II, whereas command responsibility appears in pt III, ‘General Principles of Criminal Law’. Thus, the Statutes indicate that command responsibility is a principle of liability, not an offence.
However, in order to put the strongest possible case for the ‘separate offence’ view, I can suggest a counter-argument to the structural argument. It is not unusual to have what we may call ‘adjunct’[140] offences listed in the general part of a criminal code.[141] By ‘adjunct’ offences I mean a general set of offences that map on to and are defined by reference to the specific crimes enumerated in the definitions of crimes.[142] For example, ‘attempt’ and possibly ‘incitement’ are plausibly characterised as adjunct crimes, since neither requires actual completion of the referent crime.[143] Thus, one could argue that command responsibility appears in the general part as an adjunct offence like attempts or incitement.
Nevertheless, even with this strengthened argument for a separate offence, the applicable law still seems to explicitly indicate a mode of liability and not a separate offence. For example, art 28 of the Rome Statute is quite explicit. Article 28 expressly states that the commander is held ‘criminally responsible for crimes within the jurisdiction of the Court committed by forces under his or her effective command and control’.[144] These terms seem clear that it is a ground of liability, creating liability for the core crimes committed by the subordinates.
The ICTY Statute is not as textually explicit about the nature of the liability, but there are reasons to conclude that it also recognises a mode of liability. First, art 7(3) is certainly not in the form of an offence-creating provision. Secondly, structurally, the ICTY Statute includes no adjunct offences among the ‘general principles’.[145] Thirdly, we should not lightly conclude that the ICTY Statute lays down an utterly different concept than the Rome Statute (an offence rather than a mode), to avoid unnecessary fragmentation between instruments that purport to reflect customary law.
Furthermore, the ICTY Statute purports to reflect customary law, and customary law precedent treats command responsibility as a mode of liability. It is not the aim of this article to attempt a detailed review of the doctrinal precedents, as this task has been admirably performed elsewhere.[146] For present purposes, it suffices to say that even Tribunal jurisprudence has acknowledged that command responsibility was regarded as accessory liability in the underlying crimes in post-war national legislation and jurisprudence.[147] Even the Halilović decision, in which the Trial Chamber advocated for the separate offence interpretation, shows the long consistency of the ‘mode’ approach to command responsibility. Although the Halilović decision quite bravely characterised
post-World War II jurisprudence as ‘divergent’ on the issue, every authority cited adopted the ‘mode’ approach, with the exception of only one passage from one case that arguably supported a separate dereliction offence.[148] The Halilović decision also acknowledged that national legislation treated command responsibility as a mode[149] and that the jurisprudence of the Tribunal itself had consistently done so.[150] Academic literature has generally recognised command responsibility as a mode of liability, with limited exceptions until very recently.[151]
Of course, we must allow judges scope to interpret and reinterpret provisions of their respective Statutes, especially given that ICL is a nascent discipline which is being developed each day. It is always possible that the earliest authorities did not consider all implications or did not express themselves perfectly, so that judges may later interpret them in new and better ways. As authoritative interpreters of their Statutes, judges are entitled to considerable deference. However, a judicial assertion that command responsibility has been a separate offence all along seems to run against a fair bit of history, and thus warrants at least some inquiry. Moreover, we must have some wariness where the proposal is to judicially recognise a new crime, because of the implications for the principle of legality.[152]
Thus, the option of creating a separate offence may be available to legislators, and may even be commendable. However, the solution does not appear to be available to the judges of the Tribunals or the ICC, given their current Statutes and the background of the applicable law. For the ICTY it is not an attractive solution to the culpability problem, because the approach would leap from the ‘frying pan’ of culpability problems into the ‘fire’ of legality problems. The ICTY cannot convict persons for a crime that is not in its Statute.[153]
There is a separate problem with the claim that the Tribunals do not charge the commander with the underlying crimes, namely, that the claim is demonstrably untrue. Arguments for the ‘separate offence’ approach rely on certain muddled passages in Tribunal jurisprudence,[154] but when we look at the actual charges, convictions and sentences entered by the Tribunal, we see that the Tribunals communicate to the public that the commander was a party to genocide, crimes against humanity and war crimes.
For example, claims that the commander is not charged with the underlying crime often cite a particularly infelicitously worded ‘throwaway’[155] passage in Prosecutor v Krnojelac.[156] In that passage, the Appeals Chamber stated that ‘[i]t cannot be overemphasised that, where superior responsibility is concerned, an accused is not charged with the crimes of his subordinates but with his failure to carry out his duty as a superior to exercise control’.[157] This passage is defensible insofar as it meant that the commander is not charged with committing the crimes. It is also defensible insofar as it sought to affirm that liability does not arise vicariously by virtue of the relationship but rather arises because of the commander’s personal fault: ie, her failure to exercise control.
However, if the passage is to be construed as meaning that the commander is not charged even as a party to the underlying core crime, but rather is literally charged for a distinct crime of ‘failure to exercise her duty to exercise control’, then the proposition would suffer from a major defect of not being true. The ICTY does in fact charge the commander with the crimes that were committed by subordinates. To use the example of that very case, Krnojelac was not charged with any offence of ‘failure to exercise the duty to control’. He was charged with ‘crimes against humanity and violations of the laws and customs of war’, including torture, murder, persecution, imprisonment, enslavement — ie, the core crimes that were carried out by his subordinates.[158] He was also convicted for those crimes. For example, he was found, inter alia, ‘guilty of … murder as a crime against humanity and murder as a violation of the laws or customs of war’ pursuant to art 7(3), and ‘guilty of … torture as a crime against humanity and a violation of the laws or customs of war’ pursuant to art 7(3).[159] He was not held to have ‘committed’ the crimes, but was (properly) found guilty as a party to those crimes, by virtue of command responsibility.
Other cases follow the same pattern. The practice of the ICTY shows that by virtue of command responsibility, commanders are indeed charged with the underlying crimes and sentenced as parties to the underlying crimes.[160]
Note that I am not advancing a doctrinal argument that there is ‘precedent’ for treating command responsibility as a mode. Rather, I am pointing to the empirical fact that the Tribunal demonstrably does in fact charge and convict commanders as a party to the underlying crime. It is not satisfactory to avoid the culpability concern by claiming that the commander is not held responsible as a party to the core crime, while the charges and convictions do precisely that. The charges and sentences issued by the Tribunal communicate to the public a liability in relation to crimes of genocide, crimes against humanity and war crimes. In order to impose such labels, for crimes bearing such enormous stigma, one would need to show culpability for those crimes, which requires some contribution to the crime for which one is convicted.
Another line of argument in Tribunal jurisprudence asserts not that command responsibility is a separate offence but that it is a ‘sui generis’ mode of liability, to which the contribution requirement simply does not apply.[161] This argument comes in a few variations.
Command responsibility can certainly be described as distinct if one means that it is a new species within the genus of modes of accessory liability. It differs from other modes of accessorial liability in that, for example, that it applies only in superior–subordinate relationships and features a broader mental element.[162] On a narrowly doctrinal and formalist approach, one could assert that the list of requirements for command responsibility happens not to include the contribution requirement seen in other modes. However, that doctrinal assertion does not provide a deontological justification as to why contribution is not required. It does not explain why this one mode of accessorial liability is exempt from the contribution requirement and how that is consistent with the fundamental principle of personal culpability. Such a justification would require the development of a careful principled argument.[163] Simply affixing the label ‘sui generis’ is not an argument; it does not address the culpability concern.
A more sophisticated variation of the ‘sui generis’ argument, advanced by some scholars, is that command responsibility is not merely a new mode of accessory liability but an entirely new category altogether: neither a separate offence, nor a mode of principal liability, nor a mode of accessory liability.[164] Such arguments are intriguing. One obvious attraction seems to be that, by denying that command responsibility falls within any known category (separate offence, principal liability, accessory liability), we might sidestep the deontological requirements attaching to each of the known categories. However, that still leaves a requirement to tell us what this new category is. Once the relationship between the accused and the crime in the purported new category is stated with precision, we can try to discern the appropriate deontological requirements (which would of course be non-zero). More problematically, the three categories of direct (principal) liability for the base crime, indirect (accessory) liability for the base crime, or liability for a different crime[165] seem to exhaust the logically conceivable universe of alternatives. If the claim is to be made that another category is possible, the gap should be explained and conceptually located.[166] Until then, scepticism about ill-defined new categories seems warranted. Applying Occam’s razor, it is for now more parsimonious and elegant to work with the existing categories, which appear to be mutually exclusive and jointly exhaustive.
Claims about the ‘sui generis’ nature of command responsibility have been fuelled by increasingly tortuous and convoluted statements in Tribunal jurisprudence. The jurisprudence has tied itself in knots trying to explain that there is no contradiction between a mode that does not require contribution, and ICL’s accepted principle that modes require causal contribution. The confusion is further fuelled by imprecise terminology. Consider for example the remarkable controversy over the word ‘for’. Early jurisprudence acknowledged that the commander is responsible for the crimes of the subordinates.[167] Later cases struggle to clarify that responsibility ‘for’ the crimes does not necessarily mean responsibility ‘for’ the crimes, but rather ‘because of’ the crimes, or that it is ‘not a direct responsibility’ for those acts, or that the commander does not share the ‘same responsibility’ as the perpetrators.[168] Such cases then often slip and contradict themselves again, referring again to responsibility ‘for’ the crimes of the subordinates.[169] Nonetheless, these seemingly contradictory passages could be partially defended; they are arguably just using awkward, imprecise and inconsistent terminology. As noted above, the passages are sound insofar as they are simply affirming that the commander is not treated as if she perpetrated or committed the crimes, and that is she is not held vicariously liable by virtue of the relationship but rather is liable because of her own conduct.[170]
Indeed, what the Chambers seem to be struggling to describe — that the commander is not deemed a perpetrator, that she is not held vicariously liable, and that she is held responsible for her fault in relation to the crime — is quite beautifully and simply captured by an existing concept. That existing concept is accessory liability. We do not need to fabricate an entire untested and vaguely-described conceptual category to capture these features of command responsibility.[171]
Another problem with placing command responsibility in an ill-defined twilight world between mode of liability and separate offence is that it enables a kind of ‘shell game’. The ambiguity allows the ‘mode’ nature of command responsibility to be downplayed when the culpability principle is being discussed, and then to shift back to a mode of liability just in time for conviction and sentencing. James Stewart has aptly described related arguments as ‘more of a smokescreen to ward off conceptual criticisms than a marked normative change’.[172]
Another alternative, suggested by some scholars, is a variegated account in which command responsibility operates sometimes as a mode and sometimes as a separate offence, based on variables such as failure to prevent/failure to punish, knowledge/should have known, or contributory/non-contributory.[173] Such approaches are principled, thoughtful and commendable in that they are sensitive to and compliant with the culpability principle. National legislatures are certainly free to recognise contributory derelictions as a mode and non-contributory derelictions as a separate offence.[174] My concern however is that reading such approaches into the ICL texts injects a level of complexity that is implausible on the text and that is in any event unnecessary. The relevant texts, such as art 28 of the Rome Statute, do not indicate that command responsibility operates completely differently in different instances. Article 28 gives no indication of an offence provision hidden inside it, nor does the Statute provide sentencing provisions or other rules for that offence.[175]
Most importantly, reading in a variegated approach is simply not necessary. Recall that the need for a ‘separate offence’ variation arose only because of an early decision in Tribunal jurisprudence to reject a causal contribution requirement. The variegated approach therefore became necessary to deal with non-contributory derelictions consistently with the culpability principle. The Rome Statute, by contrast, expressly requires causal contribution. It avoids the problem that set off the entire cascade of reasoning. There is no problem requiring the solution of a variegated approach. It is simply an accessory mode of liability, it requires causal contribution and hence complies with culpability, and can be applied in a straightforward manner in accordance with its text.
One seeming advantage of the variegated approach is that it allows the ICC to deal with non-contributory derelictions, whereas the approach I have outlined cannot. Consider however that the contribution requirement can be quite easily satisfied. One need only show that the failure to punish crimes increased the risk of subsequent crimes and that crimes within the ambit of that risk occurred.[176] A non-contributory dereliction arises only where there is no possibility that the dereliction could have contributed to core crimes. Consider also that the mandate of the ICC is to focus on persons most responsible for the most serious crimes of concern to the international community as a whole. I would argue that the inability of the ICC to pursue derelictions that could not have contributed to any core crimes is a non-problem. It is a non-problem that has consumed too much attention in ICL discourse. Non-contributory derelictions can be addressed by national jurisdictions, including through legislation that recognises derelictions per se as an offence in their own right. The ICC should focus on persons who might actually have at least contributed to genocide, crimes against humanity or war crimes.
In conclusion, while establishing a separate offence for command derelictions would indeed solve the culpability problem, it is doubtful, however, that this solution is available to international tribunals with their existing Statutes and applicable law. Moreover, the claim that commanders are not charged for the underlying crimes is contradicted by the actual charges and convictions issued by the Tribunals and the stigma they communicate to the public. Given that the Tribunals do indeed hold the commander liable as party to the core crimes, one cannot avoid dealing with the principle of culpability in relation to those crimes.
Further, the bare assertion that command responsibility is a ‘sui generis’ mode of liability does not furnish an answer to the culpability problem; a desert-based explanation would be required. The ‘variegated’ approaches, treating command responsibility sometimes as mode and sometimes as offence, are sophisticated but are unnecessarily complicated, because the Rome Statute avoids the entire problem by requiring causal contribution. Given the mandate of the ICC over persons most responsible for the most serious crimes, non-contributory derelictions are a non-problem that have consumed too much attention already.
The final, and most ambitious route, to question the contribution requirement would be to offer a principled, desert-based theory for a lower threshold of contribution or to dispense with contribution entirely. Whereas the foregoing Parts examined the major arguments that have dominated the debate, this Part engages in a deontological analysis of the causation issue. This Part will examine:
(a) why we require contribution;
(b) the degree of contribution required; and
(c) possible deontologically-grounded alternatives to causation for culpability.
The current article, lengthy as it is, is only able to scratch the surface of the issues raised here. The aim is to show that the arguments most commonly given to avoid the contribution requirement are insufficient and that a significant problem of fundamental principles exists. I now sketch an outline of what a principled analysis would entail. This is only a preliminary introduction, and I will draw on some key works from English language sources, which in itself presents a vast literature that I will only touch upon here. Theoretical works in other languages and from other traditions introduce some different approaches, problems and solutions. Thus, this article only introduces some of the issues and possible lines of thought, and is only a first step in a conversation to develop a cosmopolitan deontological account.[177]
As with any legal–philosophical concept, it is difficult to dig down to any unassailable bedrock as to why we require a contribution to a deed to consider a person culpable for that deed.[178] It has however been observed that the legal systems of the world seem to reflect the intuition that causation is vital and basic in assessments of responsibility.[179] It has also been argued that in a
justice-oriented system of criminal law, criminal responsibility must track moral responsibility, which in turn tracks natural relations such as causation.[180] The requirement of causal contribution matches the intuition and the principled commitment that we hold persons culpable for their acts and the blameworthy consequences thereof, and not for events in which they did not participate or to which they made no contribution. Causation is ‘the instrument we employ to ensure that responsibility is personal’.[181] As John Gardner argues, there is no way to participate in the wrongs of another other than by making a causal contribution to them.[182] Causation has been described as ‘deeply characteristic of human thought’ and expressed even among diverse societies.[183]
As discussed in Part III, accessory liability does not require ‘causation’ in a sine qua non or a ‘but for’ sense; it is sufficient to make some contribution such as encouraging, facilitating or having an effect on the crime.[184] Michael Moore writes, ‘[t]o be an accomplice, my act must have something to do with why, how or with what ease the legally prohibited result was brought about by someone else’.[185] Plausible arguments are advanced for more elastic concepts of contribution requirement, including that the conduct could have made a difference to the crime, or that it at least elevated the probability of the crime occurring.
Some arguments assert that causation should not be required for accomplice liability, but turn out on closer inspection to refer to the ‘but for’ concept of causation, and thus are simply affirming that a ‘but for’ contribution is not required for accessory liability.[186] There are more innovative arguments suggesting that even the more diluted forms of causation are not required; this is discussed below.
As discussed in Part III, ICL jurisprudence generally adheres to the principle that accessory liability requires some contribution to the underlying crime. It is ‘firmly established that for the accused to be criminally culpable his conduct must … have contributed to, or have had an effect on, the commission of the crime’.[187] Thus, the contribution may ‘make the performance of the crime possible or at least easier’.[188] The contribution must have some substantial or significant effect that furthers or facilitates the commission of the crime,[189] as long as it is prior to the full completion of the crime.[190] This approach is consistent with the most common articulation of the requirement in criminal law theory, namely that the accessory’s conduct must encourage, contribute to or facilitate the crime, render the crime easier, or put the victim at a disadvantage.[191]
I must very briefly acknowledge the debate in criminal law theory about whether omissions can be said to make a ‘contribution’ to events. It can be argued that an omission merely fails to avert the event and cannot be described as a ‘cause’ or as making a contribution.[192] While the philosophical debate over omissions and causes is obviously connected to the theme of this paper and interests me deeply, it is both unnecessary and infeasible to explore that issue in depth here. A meaningful exploration would require a separate treatise. For the purposes of this article (ie, exploring a culpability contradiction and its ramifications), it suffices to rely on the excellent responses and analyses already provided in the ample discourse on the question.[193]
Most of the scholarly literature concludes that omissions can make causal contributions, and the jurisprudence of liberal criminal justice systems has even less difficulty identifying omissions as causal contributions. For example, if a pilot aboard an aircraft has a duty to operate and land the aircraft, and yet chooses instead to do nothing and allow the plane to crash, most jurists would have little difficulty concluding that the pilot’s omission contributed to the crash. It is true that gravity and other factors would also have contributed,[194] but under common notions of causation and responsibility we would not hesitate to find that the pilot’s omission to fulfil her duty was indeed a contributing factor (indeed the major contributing factor) and that the crash was a result of her culpable inaction.[195]
The counterfactual analysis of an omission mirrors the analysis of an act. Where there was a positive act by the accused, we imagine the world where she did not do the prohibited act, to assess the difference that her act likely made. In the case of an omission, we imagine a counterfactual world where she did what was legally required, and assess the likely difference.[196] Thus, for example, the omission of a commander to establish a system of discipline, to repudiate and punish crimes, and to take appropriate steps to inculcate respect for humanitarian law, may indeed be found to encourage, facilitate or have an effect on subsequent crimes. Whether one prefers to use labels such as hypothetical causation, counterfactual causation, quasi-causation or negative causation is not of interest at this point; what matters is that there is ample plausible ground to conclude that omissions can satisfy the causal contribution requirement.
The test cited above, drawn from Tribunal jurisprudence, suggests that the accused’s conduct must be shown to have in fact made some specific contribution to the crime, however slight. However, there are plausible articulations which are broader, in that they would only require that the contribution could have made a difference, or that it at least increased the risk of the crime occurring and being successfully completed. Among the justifications for these articulations is the difficulty of tracing causal contributions to human decisions.[197] As many have noted, it is very difficult to assess the influence of one person’s conduct on the voluntary and informed acts of other human beings.[198] This is part of the reason why ‘but for’ causation is restricted to liability as a principal, whereas accessory liability requires only that the accessory’s conduct ‘contributed’ to the ultimate crimes. Some justify this more elastic standard on the grounds that influencing human behaviour is of a different nature than influencing the physical universe,[199] others would simply say that human behaviour is less predictable;[200] in either case assisting the acts of others is sufficient grounding for accessory liability.
Experience shows that it often cannot be said with certainty whether the act of a potential accomplice had a specific effect on the crime. For example, did the words of encouragement or suggestions for a plan actually have an impact on the perpetrator, or might the perpetrator have done the exact same thing in any event? If the accomplice had not provided the keys to the door, might the perpetrator have found an equally effective means of entry?[201]
In light of these difficulties, there is both jurisprudential (doctrinal) and theoretical (deontological) support for broader articulations of the contribution requirement. In a seminal article, Sanford Kadish noted that in jurisprudence, the contribution requirement is found not to be satisfied only where there was no possibility that the accomplice’s conduct could have contributed, such as where the attempted contribution demonstrably ‘never reached its target’.[202] He argued that the test — both in practice (the jurisprudence of courts) and in theory (the underlying deontological commitment) — is simply that the contribution be of a nature that ‘could have contributed to the act of the principal’.[203] Others have articulated the contribution requirement in a similar manner: that the contribution could have made a difference.[204]
In a closely related vein,[205] another articulation of the contribution requirement is that the contribution need only aggravate the risk of the crime occurring. Thus the accomplice makes a sufficient contribution when the accomplice’s conduct increases the risk of the crime occurring and successfully being completed, and a crime within the ambit of that risk occurs. Kai Ambos has noted that risk aggravation would be appropriate for culpability in command responsibility,[206] and Roberta Arnold has explored similar lines.[207] Hart and Honoré discuss the significance of risk aggravation in notions of causal contribution as used in common law and continental systems as well as everyday usage.[208] Hart and Honoré also show that in law and in common usage, a culpable dereliction in providing an opportunity for others to commit crimes can be considered a contribution to crimes that occur within the ambit of the created risk.[209]
Significantly, Tribunal jurisprudence still falls short of even these more permissive articulations of the contribution requirement, because it permits convictions even where there is no possibility that the commander’s failure contributed to core crimes or raised the risk of the crimes that occurred. As will be discussed below in Part VIII, the ICC Pre-Trial Chamber in Bemba adopted a risk aggravation theory, which seems to be supported by the theoretical work noted here.
The most promising avenue for a desert-based account that dispenses with causal contribution would be to develop a ‘ratification’ theory of culpability.[210] On a ratification theory, it would be argued that the commander, by failing to prosecute the crime, ‘ratifies’, endorses or acquiesces in the crime, and thus voluntarily absorbs liability for it ex post facto.
Such a theory would be innovative, as it goes against the principle of personal culpability as it is generally articulated and recognised, including in ICL. Nonetheless, as was mentioned in Part IV, we should be ready to re-examine and re-articulate our principles if a convincing deontological account can be developed. A preliminary exploration of this possibility has been tentatively outlined by Christopher Kutz. Acknowledging the requirement that the accomplice’s conduct at least could have contributed to the crime, he asks ‘whether … the boundaries of complicity might stretch further, to encompass acts of ratification and endorsement’.[211] Kutz suggests that perhaps we require a causal contribution only to distinguish non-culpable gestures, such as declarations of support, from types of conduct likely to enhance the risks of harm or wrongdoing. He notes that in some organisational situations, such as supervisors or lawyers providing advice, endorsement after the fact may be ‘more than an expression of approval’ but rather an act ‘with real institutional consequences’.[212] He acknowledges that ‘ratification has not been a recognised basis of criminal liability’ but argues that there are ‘good reasons to recognise it in a well-defined context’.[213]
Working along similar lines, Amy Sepinwall advances a deontological argument that the endorsement or acquiescence of a commander in failing to punish a crime might be a legitimate basis for liability in that crime.[214] She acknowledges that this approach does not fit within the formulation of the culpability principle in ICL, because the commander does not make a material contribution to the crime itself.[215] However, she advances an innovative but potentially plausible argument about culpability based on the expressive harm of endorsing crimes.[216]
A similar but slightly different approach is advanced by Guénaël Mettraux. He argues that causation is required by fundamental principles, but regards causal contribution as satisfied by the ex post enabling of impunity, ie, allowing the crime to remain unpunished.[217] While structured differently from the Sepinwall approach, this approach has similarities and produces a comparable outcome. It would constitute a departure from the contribution requirement as articulated in ICL jurisprudence and other sources (see Parts III and VII). However, one could argue for an expanded concept of culpability in which a contribution to impunity for the crime is a sufficient contribution. One might draw on ‘accessory after the fact’ as a comparable concept (see below).
There are traces of doctrinal support for a ‘ratification’ theory in command responsibility jurisprudence. For example, in Prosecutor v von Leeb (also known as the High Command case)[218] a commander was held liable for crimes prior to his failure to punish, on the grounds that his failure to take corrective action showed that he tolerated and approved of the crimes.[219] Similarly, in the Trial of
Lieutenant-General Shigeru Sawada before the Tokyo Tribunal, a commander learned of crimes that occurred while he was away and failed to punish those responsible. The Tribunal held that by his failure to punish he ‘ratified’ the crimes.[220] An ICTY Trial Chamber has described the ‘tacit acceptance’ argument as ‘not without merit’, because a failure to punish is ‘so grave that international law imposes upon him responsibility for those crimes’.[221]
Thus there may be some doctrinal precedent in ICL for such a position. Doctrinal precedent does not however answer our principled question, which is: can such an approach be justified? Does it accord with our principles?
In developing such an argument, one might explore the following avenues:
1 the concept of ‘accessory after the fact’;2 the state responsibility doctrine of ‘adoption’;
3 concepts of ‘pollution’; and
4 the role of will and expression.
A commander’s failure to punish a crime may not contribute to that crime (given that the crime has already happened), but it does facilitate the perpetrator’s escape from justice and helps allow the crime to go unpunished. One could highlight the similarity to the concept of ‘accessory after the fact’, a concept once familiar in common law systems, or the comparable concept of ‘assistance après coup’ or facilitation after the fact once known in some civil law systems.[222] The early concept was that a person assisting a perpetrator to avoid arrest, trial or conviction thereby becomes party to the original felony.[223] The parallel with command responsibility is clear, because failing to meet the duty to punish clearly assists the perpetrators to avoid arrest, trial or conviction. One could argue that if ‘accessory after the fact’ is justifiable under the culpability principle, then a concept of command responsibility without causation is also justifiable, because the commander also contributes to the frustration of justice.
The difficulty with this analogy is that liberal systems have moved away from treating assistance after the fact as creating liability for the crime itself, precisely because of concerns with culpability.[224] The contemporary approach is to punish aiders after the fact for what they actually did, such as obstructing justice or harbouring a fugitive. We do not hold them retroactively responsible for an already-completed crime in which they did not participate and to which they did not contribute, because it was not in any way their deed, even on the broader articulations of the contribution requirement.[225] Thus, post-offence assistance in frustrating justice is no longer seen as a mode of liability in the crime but as a separate crime. As noted in Part VI, however, Tribunal practice explicitly holds the commander guilty as party to the core crime. Thus, analogy to accessory after the fact does not provide a principled basis to treat non-contributory derelictions as a mode of liability.
An argument could also be attempted by drawing on the public international law concept of state responsibility by ‘acknowledgement and adoption’. Under this doctrine, a state may acknowledge an act, carried out by persons who were not its agents, and ‘adopt’ that conduct ex post facto as its own conduct, thereby acquiring liability for that conduct.[226]
However, this analogy also raises difficulties. State responsibility operates in a field of delicts, creating civil liability (for example, a requirement to pay compensation). It is a different question whether an individual may endorse the conduct of other persons after the fact and thereby become criminally complicit in it. For example, if one individual throws a shoe at a president, there may be others who approve of the act, but no matter how hearty their approval and how fervent their desire to make it ‘their’ act, they cannot voluntarily assume criminal liability for the act. It is not their act. Thus, as Sayre has written, the concept of ‘ratification’ exists in private law but not in criminal law, where responsibility is personal and thus we need a contribution to the deed.[227] Similarly, Kadish notes that in criminal law, literal consent to be criminally liable is irrelevant; we need an action that furthers the crime charged.[228] Other scholars have reached the same conclusion.[229]
Another possible avenue would be to appeal to a concept of crime as ‘pollution’. On this view, the unpunished crime is a stain which pollutes the moral order, and thus requires ‘cleansing or expatiation’, which restores the law.[230] By failing to punish, the commander perpetuates the stain; the ‘bloodguilt’ comes to rest with her.[231] It is arguable that a sense of moral pollution may have undergirded early concepts of accessory after the fact or assistance après coup. This line of thought is intriguing and potentially of assistance. However, while the non-contributing dereliction of the commander may give rise to moral or metaphysical guilt for the deed, it may not be an adequate basis to assign criminal guilt for the deed.[232] The idea of bloodguilt and of moral pollution tainting other actors may have a role in assessing broader concepts of responsibility,[233] but it is arguably a rather poetic or metaphorical basis for a contemporary and rational system of justice to assign criminal guilt.[234] Without ruling out a ‘pollution’-based approach, it may be said for now that it is not clear that this can provide a reliable deontological justification to ascribe criminal culpability for the crime.
The most convincing justification for a ratification theory of culpability would be to link the crime to the expression of will of the accused. One could argue that contributions to an act render the accessory culpable because they are manifestations of the accomplice’s will in relation to that act.[235] It could then be argued that what truly underscores complicity is the manifestation of will in relation to the crime, and that contributions are merely one means of manifesting the accessory’s will. The concept of will could then be extended further, so that a commander’s ex post facto failure to punish might be regarded as extending her will to encompass the crime, and thereby to create liability in relation to it. This approach is arguably compatible with the suggestion of Amy Sepinwall that, from the perspective of the victim, the endorsement or acquiescence by the commander contributes to the expressive harm of the crime and generates liability.[236]
A will (or expression) based approach to culpability is intriguing, and would require some significant rethinking of the culpability principle. It is a valuable line of inquiry. There are some additional difficulties that would have to be resolved before such a theory could provide a solution to the command responsibility question. One set of concerns to be addressed would be that culpability based on expressions of will may be a rather elastic approach to liability; it lacks the concreteness of requiring a causal contribution in order for a deed to be attributed to the accused. There is also a second hurdle in relation to command responsibility in particular. While it may be plausible to speak of extending one’s will to the deed where the commander at least has actual knowledge, command responsibility can be based on mental elements significantly more inclusive than actual knowledge. Thus a theory based on will may not be suitable to explain and justify indirect liability without causal contribution and without subjective knowledge.[237]
In conclusion, a ‘ratification’ theory may be the most plausible avenue to justify liability without causal contributions. Such an account certainly deviates from the culpability principle as it is widely understood and as it is conceived in ICL, and thus would require careful justification. Some thoughtful writers have begun to explore this possibility. While views may reasonably differ, the discussion is a welcome advance on purely doctrinal and instrumentalist debates, because it acknowledges and grapples with the system’s commitment to respect individuals as moral agents and the need for some deontological justification. For the time being it seems that there are still difficulties to be overcome before ratification theory can offer an explanation and justification for command responsibility as a mode of liability without causal contribution.
At present, the dominant approach in Tribunal jurisprudence regards command responsibility as a mode of liability and yet rejects a causal contribution requirement. This article has sought to demonstrate that this is an extraordinary position. Tribunal jurisprudence declares its respect for personal culpability, which means that persons may be held responsible only for crimes in which they have participated or to which they have contributed.[238] Thus, there appears to be an internal contradiction between the doctrine and the stated principle. This contradiction has been obscured by many factors, including persistent ambiguity about the nature of command responsibility.
The jurisprudence has not seriously acknowledged or grappled with this contradiction. Technical doctrinal arguments have been advanced in support of the rejection of causation; however, those arguments are not sustained even by their own doctrinal premises. Moreover, they do not even attempt to address the fundamental deontological culpability concern, which is rooted in the permissible limits of punishment and the ‘foundation of criminal responsibility’.[239]
Illumination of the contribution requirement of personal culpability also sheds new light on the controversy over successor commander liability. The majority position in Hadžihasanović, which required that the commander be in command at the time of the crimes, has been heavily criticised. It appears from the Orić appeal decision that a new majority of appellate judges are inclined to reverse the decision as an unfortunate and arbitrary limitation. The current debate largely centres on doctrinal questions about what the precedents permit and on teleological arguments about maximising the impact on commanders. What is largely missing from the debate is that the idea that holding a person liable for crimes that were completed before she even joined a unit may be a remarkably unjust thing to do. It is a significant departure from the principle of culpability as articulated by the system itself and as currently understood. Thus, to sustain the innovation of successor commander liability would require a compelling justification and a new understanding of culpability.
The Hadžihasanović decision creates an unexpected ‘intermediate’ position with respect to causal contribution in command responsibility. The two most obvious options would have been to either require causal contribution, as the Rome Statute does, or to reject causation and thus to embrace successor commander liability. Instead, under Hadžihasanović, Tribunal jurisprudence rejects a contribution requirement but requires that the commander was at least in command at the time of the crimes. By requiring contemporaneous command, the Hadžihasanović majority position reduces much of the culpability deficit. The requirement of contemporaneous command means that in most instances the dereliction would at least satisfy the broader articulations of the culpability requirement, such as that the commander’s failures ‘could have’ contributed to the crimes or that they elevated the risk of criminality. The majority approach therefore screens out at least the most egregiously inappropriate cases.
Nonetheless, the Tribunal jurisprudence is still potentially problematic. There may be instances where a commander was in charge at the time of the crime or crimes and yet her derelictions could not possibly have contributed to core crimes. One example is the ‘isolated crime’ scenario, where the only fault of the commander is that she failed to punish past crimes, but no further crimes occur.[240] In this scenario, her dereliction obviously did not retroactively facilitate the crime that occurred, nor were there any subsequent crimes that her failure may have encouraged or facilitated.[241] Yet Tribunal jurisprudence would allow her to be convicted of core crimes by virtue of command responsibility. This liability without contribution contradicts the principle of culpability as stated by the system. There is no core crime that can be even derivatively attributed to her.
The simplest solution would be to recognise that command responsibility does indeed require that the commander’s failures facilitated or encouraged crimes, or at least that the failures increased the risk of crimes and that crimes within the ambit of that risk occurred. There are also three alternative solutions. One would be to develop a novel ratification theory of culpability by which the failure to punish entails an absorption of liability. Another would be to establish a distinct dereliction offence through appropriate law-creating methods. The remaining alternative is to abandon the claim that the system conforms to liberal principles, and to accept any loss of legitimacy that flows from this acknowledgement.
The core implication of this article is that the ICC need not and should not import many of the more complex and convoluted claims about command responsibility that have been generated in Tribunal discourse. I have argued that early Tribunal jurisprudence adopted a fallacious argument that the ‘failure to prevent’ branch is irreconcilable with a contribution requirement, leading it to reject a contribution requirement. Subsequent efforts to reconcile this with the contribution requirement have generated many inconsistent and mysteriously vague statements about the nature of command responsibility. It has also generated suggestions that command responsibility has a variegated nature, operating sometimes as mode and sometimes as offence. Those arguments have the virtue of being less mysterious, because they rely on known concepts of liability, but they inject a complexity that the text of art 28 does not support. Moreover, a variegated approach is only necessitated by the Tribunal insistence on including non-contributory derelictions.[242] If the ICC adheres to the express contribution requirement then it can operate much more simply as a mode of accessory liability, properly requiring causal contribution.[243]
The ICC is on a different path, because art 28 of the Rome Statute expressly requires causal contribution, ie, that the crimes occurred ‘as a result of the failure of the commander to exercise control properly’. This provision has attracted considerable criticism in the literature on the ground that it is an unfortunate limitation that will hamper prosecutions.[244] However, the contribution requirement in the Rome Statute is arguably to be commended rather than condemned. As was discussed above, the contribution requirement not only has stronger doctrinal support than is commonly supposed,[245] it appears necessary to respect liberal principles.[246]
The first consideration of art 28 by an ICC Chamber was advanced by Pre-Trial Chamber II in the confirmation of charges decision in the Bemba case.[247] The Chamber’s reasoning was careful and plausible, and in particular it offered a sophisticated interpretation of the contribution requirement.
While the Chamber did not delve into the ‘mode-versus-offence’ debate, it appeared to treat command responsibility as a mode of liability, which seems the more plausible interpretation.[248] The Chamber referred repeatedly to the commander’s ‘responsibility for the crimes committed by his forces’, ie, for the underlying crimes.[249] The counts do not charge Bemba with a separate offence of failure to exercise control; they charge him with responsibility for crimes against humanity and war crimes, by virtue of command responsibility.[250] Several features of the decision indicate that command responsibility is understood as a mode of liability.[251]
With respect to the problem of the successor commander, the Chamber noted the controversy in other tribunals as to whether the commander must have been in command at the time of the crimes.[252] The Chamber sided with the majority position of the ICTY Appeals Chamber in Hadžihasanović, holding that the commander must have been in command at the time of the crimes.[253] For the reasons outlined above, this position helps ensure conformity with the culpability principle.
The Chamber also noted that art 28 expressly requires causal contribution, ie, that the crimes be ‘a result of’ a failure by the commander ‘to exercise control properly’.[254] The Chamber adopted the only feasible interpretation of ‘failure to exercise control properly’, which is that it must refer back to the basic duties to prevent or punish,[255] and that it does not require an entirely separate type of dereliction.
As for the requisite extent of the contribution, the Chamber provided a sophisticated and plausible interpretation. The Chamber considered the possibility of adopting a ‘but for’ test, ie, but for the commander’s dereliction, the crimes would not have happened. The Chamber concluded that it is only required that the commander’s omission ‘increased the risk of the commission of the crimes charged’.[256] The approach adopted is consistent with the underpinning principle of culpability: as discussed above, ‘risk aggravation’ is one of the plausible articulations of the contribution requirement.[257] The reason given by the Chamber was that assessing the impact of an omission may be more difficult than assessing that of an act,[258] whereas a stronger reason would have been the special issues in tracing influences on human behaviour.[259] In any event, the standard adopted seems to be compatible with fundamental principles.
One could question whether the ‘risk aggravation’ interpretation is compatible with the wording of the Rome Statute, but ultimately it is plausible for several reasons. On one hand, an ordinary meaning textual interpretation of the terms ‘as a result of’ may seem to connote a more stringent standard (such as ‘but for’). However, a contextual and teleological interpretation supports the Chamber’s conclusion. A plausible approach to teleological interpretation could consider that the object and purpose of the provision includes ‘victim protection, subject to fundamental principles of a liberal justice system’.[260] The risk aggravation approach serves to make the provision effective while complying with fundamental principles. Furthermore, a contextual interpretation allows interpretation in light of applicable law, and hence consideration of customary international law and general principles from the legal systems of the world.[261] Again, the risk aggravation approach to accomplice liability is well known in state practice and in general principles of criminal justice, making it a plausible interpretation.
Conversely, a ‘but for’ interpretation would provide a hopelessly stringent and ineffective test, because of the extreme difficulty or impossibility of identifying the determinants of human behaviour with that degree of certainty.[262] Nor are there principled reasons that would compel the adoption of such a stringent test, as a ‘but for’ standard is required neither by customary law, nor by legal systems of the world, nor by the principle of culpability. Accordingly, the
risk-aggravation interpretation seems textually defensible and compliant with fundamental liberal principles.
While the foregoing aspects of the Chamber’s decision are commendable, there was a potentially unfortunate passage in which the Chamber echoed some problematic reasoning from Tribunal jurisprudence. The Chamber held that it is ‘illogical to conclude that a failure [to punish] can retroactively cause the crimes to be committed’[263] and therefore asserted that the contribution requirement ‘only relates to the commander’s duty to prevent’.[264]
If the passage means that commanders may be convicted based on failures to punish without a contribution to the crimes, it is problematic. It is contrary to the explicit requirement in art 28 of a causal contribution. Of course an interpreter can adopt a counter-textual interpretation in some circumstances, but in this case the Chamber relied on the ‘incompatibility’ argument which, as was demonstrated in Parts II and V, is unsound. Moreover, rejecting the contribution requirement contradicts the culpability principle and would draw the ICC into the cascade of reasoning that has mired command responsibility in complexity, contradiction and vagueness.
The express requirement in art 28 of a causal contribution is easily reconciled with the ‘failure to punish’ branch. By failing to punish past crimes and thus failing to send a signal of disapproval and deterrence, the commander culpably elevates the risk of subsequent crimes,[265] and thus she joins in liability for any subsequent crimes falling within the ambit of the created risk. This satisfies all terms of art 28.[266] There is nothing ‘illogical’ or self-contradictory that would require us, or even permit us, to disregard a requirement that is expressly stated in the Rome Statute and which is also required by the culpability principle.
The text of art 28 requires a contribution to crimes regardless of whether liability is based on failures to prevent or failures to punish. The two branches provide two routes to establish liability: a prosecutor may show either that a failure to take reasonable preventive steps facilitated crimes, or that a failure to investigate and punish prior crimes facilitated subsequent crimes. Either failure by a commander is sufficient to provide grounding for accessory liability.
Perhaps the best reading is to downplay the problematic passage. Indeed, the Chamber nonetheless went on to note (correctly) that a failure to punish past crimes ‘can have a causal impact on the commission of further crimes’ and ‘is likely to increase the risk that further crimes will be committed in the future’.[267] With this in mind, we can respect the terms of the Rome Statute and the fundamental principle of culpability.
The core of this article is the following syllogism:
1 ICL claims to be a liberal system of criminal justice.2 A liberal system of criminal justice includes the principle of personal culpability.
3 The principle of personal culpability requires that persons only be held liable for crimes to which they contributed.
4 The doctrine of command responsibility holds the commander liable as a party to the crimes of the subordinates.
5 Therefore, to comply with our stated principles, command responsibility must require that commander’s dereliction contributed to the crimes of subordinates.
I have sought to show that there is therefore a problem in the ICTY jurisprudence and its rejection of causal contribution. The problem sits in plain sight, but has been partially obscured in the jurisprudence and discourse by various ambiguities and controversies. I have therefore examined each argument in the hopes of bringing the problem into clearer relief.
For the Tribunals and entities embracing Tribunal jurisprudence, the most obvious solution is to bring the doctrine into conformity with the stated principle. The alternatives are: to reject liberal principles (which seems undesirable), to posit command responsibility as a separate offence (which seems legally unavailable to the Tribunals) or to advance an innovative deontologically-justified theory of culpability (which requires more development).
In terms of reasoning, I have sought to show that legal reasoning has often failed to grapple with fundamental principles. Tribunal jurisprudence took an early misstep in rejecting causal contribution for inadequate reasons. This put the jurisprudence and discourse onto a particular path. Rather than reviewing the first misstep, an increasingly complex edifice has been constructed to try to explain and justify it. For the ICC, where causal contribution is expressly required, all of those pitfalls and resulting claims can simply be avoided. The Tribunals have generated considerable useful jurisprudence and have been valuable laboratories, but we can learn from both the successful and unsuccessful experiments; the approach to causal contribution is one of the latter.
In terms of doctrine, I have sought to show that the ‘failure to prevent’ branch is quite compatible with the contribution requirement. The belief that it was not (the incompatibility argument) took the jurisprudence and some of the literature away from a simple and principled solution. Command responsibility in customary international law is, quite simply, a mode of accessory liability that requires a causal contribution to the principal crimes. National legislators may of course supplement this with new offences capturing derelictions, including derelictions that do not contribute to crimes.
Finally, in terms of theory, there is principled support for broad approaches to the contribution requirement, including a ‘risk aggravation’ standard. Thus the approach adopted in the Bemba confirmation decision appears to be a plausible articulation of the contribution requirement. In the context of failures to prevent, it should in most imaginable cases be straightforward to show that a failure to take adequate steps to prevent crimes increased the risk of such crimes.[268] As for failures to punish, a failure of the commander to punish and repudiate crimes can facilitate, encourage or increase the risk of subsequent crimes, and can ground accessory liability for those subsequent crimes. In circumstances where there is no possibility that the commander’s omissions contributed to crimes, then liability for those crimes is inappropriate. Moreover, international institutions mandated to deal with the persons most responsible for the most serious crimes need not be pre-occupied with derelictions that did not contribute to even a single core crime. The commander may be liable under national laws for her dereliction of duty, but we should not treat her as party to serious international crimes in which she is not implicated.
[*] Queen’s University, Faculty of Law. Research for this project was facilitated by funding from the Social Sciences and Humanities Research Council of Canada. I am deeply grateful for the valuable comments provided by Kai Ambos, Mark Drumbl, Malcolm Thorburn and the anonymous peer reviewers, as well as the excellent research assistance of Deborah Bayley.
[1] ‘Tribunal jurisprudence’ refers to case law from the International Criminal Tribunal for the Former Yugoslavia and the International Criminal Tribunal for Rwanda. Citations of cases heard before these Tribunals will adopt the conventional acronyms ICTY and ICTR, respectively. Similarly, ‘ICC’ will denote cases heard by the International Criminal Court. The establishing statutes of each of these bodies will be collectively referred to as the ‘Statutes’.
[2] Rome Statute of the International Criminal Court, opened for signature 17 July 1998, 2187 UNTS 90 (entered into force 1 July 2002) (‘Rome Statute’).
[3] This premise may be particularly controversial for many readers, but I will elaborate below in Part VI.
[4] The problem with the analogy to ‘accessory after the fact’ is that the idea of holding the post-offence aider as an actual party to the completed crime has been widely abandoned as unprincipled in liberal systems. Instead the contemporary approach generally holds the aider responsible for what she has actually done (eg, obstruction of justice). Thus, this particular avenue is not of assistance given that ICL institutions hold the commander responsible as a party.
[5] See, eg, Amy J Sepinwall, ‘Failures to Punish: Command Responsibility in Domestic and International Law’ (2009) 30 Michigan Journal of International Law 251.
[6] Prosecutor v Hadžihasanović (Decision on Interlocutory Appeal Challenging Jurisdiction in relation to Command Responsibility) (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) (‘Hadžihasanović’).
[7] Prosecutor v Bemba (Decision pursuant to Article 61(7)(a) and (b) of the Rome Statute on the Charges of the Prosecutor against Jean-Pierre Bemba Gombo) (ICC, Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009) (‘Bemba’).
[9] The Anglo-American usage of the term ‘doctrine’ is more akin to what other traditions might call ‘positive law’.
[10] SC Res 827, UN SCOR, 48th sess, 3217th mtg, UN Doc S/RES/827 (25 May 1993) art 7(3) (‘ICTY Statute’); SC Res 955, UN SCOR, 49th sess, 3453rd mtg, UN Doc S/RES/955 (8 November 1994) art 6(3) (‘ICTR Statute’); Prosecutor v Kordić (Judgement) (ICTY, Appeals Chamber, Case No IT-95-14/2-A, 17 December 2004) [839].
[11] Article 28 of the Rome Statute provides that:
In addition to other grounds of criminal responsibility under this Statute for crimes within the jurisdiction of the Court:
(a) A military commander or person effectively acting as a military commander shall be criminally responsible for crimes within the jurisdiction of the Court committed by forces under his or her effective command and control, or effective authority and control as the case may be, as a result of his or her failure to exercise control properly over such forces, where:
(i) That military commander or person either knew or, owing to the circumstances at the time, should have known that the forces were committing or about to commit such crimes; and
(ii) That military commander or person failed to 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.
[12] Article 28(b) of the Rome Statute contains a provision for civilian superiors that is largely similar to art 28(a). However the requisite mental element is that the civilian superior ‘knew, or consciously disregarded information which clearly indicated, that the subordinates were committing or about to commit such crimes’: art 28(b)(i).
[13] Other complicity doctrines include ‘joint criminal enterprise’ before the Tribunals and contribution to a ‘common purpose’ before the ICC: see, eg, Rome Statute art 25.
[14] Prosecutor v Orić (Judgement) (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [283] (‘Orić’); Prosecutor v Orić (Judgement) (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) [43] (‘Orić Appeal’); Prosecutor v Kvočka (Judgement) (ICTY, Appeals Chamber, Case No IT-98-30/1-A, 28 February 2005) [187]. For a similar approach in the common law, see A P Simester and G R Sullivan, Criminal Law: Theory and Doctrine (Hart, 3rd ed, 2007) 204–7; Andrew Ashworth, Principles of Criminal Law (Oxford University Press, 5th ed, 2006) 410.
[15] Beyond the technical value-added, command responsibility, of course, also has a pedagogic function as a label and signal to commanders as to their potential liabilities. Thus, cases of actual knowledge, which could technically be prosecuted as aiding and abetting by omission, can be prosecuted under command responsibility as they satisfy its terms, and doing so reinforces the didactic function of command responsibility doctrine.
[16] See, eg, Mirjan Damaška, ‘The Shadow Side of Command Responsibility’ (2001) 49 American Journal of Comparative Law 455; Alison Marston Danner and Jenny S Martinez, ‘Guilty Associations: Joint Criminal Enterprise, Command Responsibility, and the Development of International Criminal Law’ (2005) 93 California Law Review 75, 150–1; Arthur T O’Reilly, ‘Command Responsibility: A Call to Realign the Doctrine with Principles of Individual Accountability and Retributive Justice’ (2004) 40 Gonzaga Law Review 127; David L Nersessian, ‘Whoops, I Committed Genocide! The Anomaly of Constructive Liability for Serious International Crimes’ (2006) 30 Fletcher Forum of World Affairs 81; Ann B Ching, ‘Evolution of the Command Responsibility Doctrine in light of the Celebici Decision of the International Criminal Tribunal for the Former Yugoslavia’ (1999) 25 North Carolina Journal of International Law and Commercial Regulation 167, 204; Shane Darcy, ‘Imputed Criminal Liability and the Goals of International Justice’ (2007) 20 Leiden Journal of International Law 377.
[17] See, eg, Michael Duttwiler, ‘Liability for Omissions in International Criminal Law’ (2008) 6 International Criminal Law Review 1.
[18] As will be discussed below in Part VII, however, the requirement of personal dereliction may not completely address the ‘vicarious liability’ concern. The principle of culpability requires not only fault, but fault for the crime charged; in the absence of causal contribution, it is arguable that the doctrine falls short of the principle.
[19] SC Res 827, UN SCOR, 48th sess, 3217th mtg, UN Doc S/RES/827 (25 May 1993), as amended by SC Res 1877, UN SCOR, 64th sess, 6155th mtg, UN Doc S/RES/1877 (7 July 2009).
[20] Prosecutor v Blaškić (Judgement) (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [62] (‘Blaškić’): describing the ‘had reason to know’ standard and asserting that it is not criminal negligence; Prosecutor v Delalić (Judgement) (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998) (‘Čelebići’); Bemba (ICC, ICC-01/05-01/08, 15 June 2009) [427]–[434]: describing the ‘should have known’ standard as one of negligence. See Guénaël Mettraux, The Law of Command Responsibility (Oxford University Press, 2009) 194–223.
[21] ICTY Statute art 7(1); ICTR Statute art 6(1); Rome Statute art 25; Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [288]; Prosecutor v Tadić (Judgement) (ICTY, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) [229] (‘Tadić’); Prosecutor v Furundžija (Judgement) (ICTY, Trial Chamber, Case No IT-95-17/1-T, 10 December 1998) [245]; Prosecutor v Simić (Judgement) (ICTY, Appeals Chamber, Case No IT-95-9-A, 28 November 2006) [86]; Sanford H Kadish, ‘Complicity, Cause and Blame: A Study in the Interpretation of Doctrine’ (1985) 73 California Law Review 323, 346; Markus D Dubber, ‘Criminalizing Complicity: A Comparative Analysis’ (2007) 5 Journal of International Criminal Justice 977, 978.
[22] Paul H Robinson, ‘Imputed Criminal Liability’ (1984) 93 Yale Law Journal 609.
[23] Examples include doctrines of complicity, acting through an innocent agent and self-induced intoxication: ibid 611.
[24] Ibid 641–9.
[25] Kadish notes that normal principles require intent, but that it is ‘not completely clear that this requirement is entailed in the core supposition of complicity doctrine and thus he left for further reflection issues of recklessness and negligence: Kadish, above n 21, 406. See also at 355, 409.
[26] Jenny S Martinez, ‘Understanding Mens Rea in Command Responsibility: From Yamashita to Blaškić and Beyond’ (2007) 5 Journal of International Criminal Justice 638. I outlined, but did not develop, a similar argument: Darryl Robinson, ‘The Two Liberalisms of International Criminal Law’ in Carsten Stahn and Larissa van den Herik (eds), Future Perspectives on International Criminal Justice (T M C Asser Press, 2010) 115, 126.
[27] As noted by Martinez, this includes a legal privilege for acts that would normally constitute crimes, as well as ‘licence to turn ordinary men into lethally destructive, and legally privileged, soldiers’: Martinez, above n 26, 661–3.
[28] See Paul Robinson, above n 22, on substitution of elements. See also David Luban, ‘Contrived Ignorance’ (1999) 87 Georgetown Law Journal 957; On the need for individual responsibility in systems of deniability, see David Luban, ‘Moral Responsibility in the Age of Bureaucracy’ (1992) 90 Michigan Law Review 2348, 2382–3. On the sufficiency of negligence for criminal liability, see, eg, H L A Hart, Punishment and Responsibility: Essays in the Philosophy of Law (Oxford University Press, 2nd ed, 2008) 132–4; George P Fletcher, Basic Concepts of Criminal Law (Oxford University Press, 1998) 111–20.
[29] This distinction between a doctrine and the underlying deontological commitment is explained in more detail in my forthcoming article, ‘A Cosmopolitan Liberal Account of International Criminal Law’.
[30] This analysis also gives reason to question the assumption, often made in the literature, that the ‘should have known’ shortcut should also be applied to civilian superiors. The Rome Statute accorded a more generous standard for civilians, which is often criticised by commentators as an unfortunate concession to self-interest: see Darryl Robinson, ‘The Identity Crisis of International Criminal Law’ (2008) 21 Leiden Journal of International Law 925, 956–61. The concern may of course be correct. Alternatively, it may be that there is a principled basis for the distinction between military and civilian superiors. The ‘equivalency’ justification outlined above pertains to the context of military commanders, who oversee trained, armed and dangerous forces, in a system of military discipline and around-the-clock supervision. It may or may not be possible to extend the justification to civilian superiors. An astute question would be why I conclude that the mens rea departure is justified but the abandonment of causal contribution is not. The answer is that the former departure, modifying one element, can arguably be given a deontological justification, for the reasons just given. Conversely, the rejection of causal contribution, including even the most elastic conceptions thereof, and hence the departure from culpability as hitherto understood, has not yet been given a deontological justification. Some efforts have been made, of which the most admirable and potentially plausible is that advanced by Amy Sepinwall, above n 5, discussed below in Part VII.
[31] On some accounts, we hold the person responsible not for the consequences per se but rather for the risks and dangers created by the conduct.
[32] As noted by John Gardner, I am responsible for my actions and you are responsible for your actions. However, my actions may influence yours, so personal culpability does not mean we have no regard to the actions of others: John Gardner, ‘Complicity and Causality’ (2007) 1 Criminal Law and Philosophy 127, 132.
[33] Mettraux, above n 20, 82. Mettraux maintains, however, that causal contribution can be satisfied by contributing to impunity for the crime: at 43, 80. This position differs from the generally recognised conception of culpability, which requires a contribution to the crime itself, and is reminiscent of earlier doctrines such as ‘accessory after the fact’. The position is discussed in Part VII.
[34] Ilias Bantekas, ‘On Stretching the Boundaries of Responsible Command’ (2009) 7 Journal of International Criminal Justice 1197, 1199. Works on causation from criminal legal theory are discussed below in Part VII. Part VII will also discuss the possibility of non-causal theories of culpability.
[35] Ilias Bantekas, ‘The Contemporary Law of Superior Responsibility’ (1999) 93 American Journal of International Law 573, 577; Kadish, above n 21, 337–42; Simester and Sullivan, above n 14, 193–6.
[36] See, eg, Héctor Olásolo, ‘Developments in the Distinction between Principal and Accessory Liability in light of the First Case Law of the International Criminal Court’ in Carsten Stahn and Göran Sluiter (eds), The Emerging Practice of the International Criminal Court (Martinus Nijhoff, 2009) 339. For an argument for a unitary model, see James Stewart, ‘The End of “Modes of Liability” for International Crimes’ (2011) 25 Leiden Journal of International Law 165.
[37] There are different possible ways to distinguish between accessories and principals; for present purposes we focus on the essential contribution, which has support in ICL jurisprudence and ICL literature: see, eg, Prosecutor v Katanga (Decision on Confirmation of Charges) (ICC, Pre-Trial Chamber, Case No ICC-01/04-01/07, 30 September 2008) [480]–[486]; Prosecutor v Lubanga (Decision on the Confirmation of Charges) (ICC, Pre-Trial Chamber I, Case No ICC-01/04-01/06, 29 January 2007) [322]–[340]. See also Olásolo, above n 36; Dubber, above n 21. In a seminal article, Kadish explains how principals make a sine qua non (but for) contribution, whereas the accomplice aids or influences the principal; the consequence of her act is the influence on the choices and actions of others: Kadish, above n 21. See also Michael S Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’ (2007) 156 University of Pennsylvania Law Review 395, 401; Joshua Dressler, ‘Reassessing the Theoretical Underpinnings of Accomplice Liability: New Solutions to an Old Problem’ (1985) 37 Hastings Law Journal 91, 99–102.
[38] Other principals (such as those acting through an innocent agent or co-perpetrating) still have ‘control’ over the crimes and make a ‘but for’ contribution: see sources cited in ibid.
[39] Kadish, above n 21, 328, 343–6; Dressler, above n 37, 139.
[40] As Gardner explains, ‘[b]oth principals and accomplices make a difference, change the world, have an influence … accomplices make their difference through principals, in other words, by making a difference to the difference that principals make’: Gardner, above n 32, 128.
[41] Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 401.
[42] See Part VII.
[43] On compliance with fundamental principles and the highest legal standards, see Counsel Support Section, ICC — Defence, International Criminal Court <http://www.icc-cpi.int/Menus/ICC/Structure+of+the+Court/Defence/> Report of the Secretary-General pursuant to Paragraph 2 of Security Council Resolution 808 (1993), UN SCOR, 48th sess, UN Doc S/25704 (3 May 1993) [34], [106]; Prosecutor v Tadić (Decision on the Defence Motion for Interlocutory Appeal on Jurisdiction) (ICTY, Appeals Chamber, Case No IT-94-1-A, 2 October 1995) [42], [45], [62]; Jelena Pejic, ‘The International Criminal Court Statute: An Appraisal of the Rome Package’ (2000) 34 International Lawyer 65, 69.
[44] (Judgement) (ICTY, Appeals Chamber, Case No IT-94-1-A, 15 July 1999).
[45] Ibid [186]. See also International Military Tribunal (Nuremberg) Judgment and Sentences, reproduced in ‘Judicial Decisions’ (1946) 41 American Journal of International Law 172, 251: ‘criminal guilt is personal’.
[46] Prosecutor v Kayishema (Judgement) (ICTR, Trial Chamber II, Case No ICTR-95-1, 21 May 1999) [199] (‘Kayishema’).
[47] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [280].
[48] Tribunal jurisprudence indicates that the only ‘exception’, in which conduct after the crime can be regarded as contributing to the commission of the crime, is where there is a prior agreement to subsequently aid or abet: Prosecutor v Blagojević (Judgement) (ICTY, Trial Chamber I, Case No IT-02-60-T, 17 January 2005) [731]. However, this is not really an exception, given that there is a prior agreement, and it is the agreement that can facilitate, encourage or have an effect on the crime.
[49] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [280]–[288].
[50] Ibid [282].
[51] Ibid.
[52] Ibid [283].
[53] Ibid [284].
[54] Ibid [338].
[55] The ICTY Statute and ICTR Statute refer to ‘failures to prevent’ and ‘failures to punish’. The Rome Statute actually splits the possible derelictions into three categories: failures to prevent, to repress, and to submit the matter to other authorities for punishment. While the three-pronged ICC approach may be useful for highlighting different obligations of commanders, this article will, for simplicity, continue to refer to the two conceptually different stages: ‘failures to prevent’ (referring to actions required prior to a particular crime) and ‘failures to punish’ (referring to actions required after a particular crime). As will be discussed below in Part VIII, the three options in the Rome Statute ultimately collapse into one of these two conceptual categories: see ICTY Statute art 7(3); ICTR Statute art 6(3); Rome Statute art 28.
[56] ICTY Statute art 7(3); ICTR Statute art 6(3). A similar requirement appears in Rome Statute arts 28(a)(ii), 28(b)(iii). The obligation is one of means and not results; the mere fact that crimes nonetheless occurred does not mean that the commander failed to meet her duty to take reasonable preventive steps.
[57] Some may argue that a failure to prevent, as an omission, cannot be regarded as ‘contributing’ to any events. This argument is discussed below in Part VII.
[58] ICTY Statute art 7(3); ICTR Statute art 6(3); Rome Statute arts 28(a)(ii), 28(b)(iii).
[59] Blaškić (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [83].
[60] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [338].
[61] Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998) [400]. The Prosecution similarly rejected the possibility of causal nexus ‘as a matter of logic’: at [397].
[62] Ibid [396]–[400], endorsed in Blaškić (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [76].
[63] I use the term ‘criminal negligence’, following the ICC approach, to avoid verbiage. The standard is, of course, ‘had reason to know’ in Tribunal jurisprudence, which has avoided reliance on the concept of criminal negligence.
[64] Another objection that has been offered is that ‘aiding and abetting’ by omission already covers such scenarios. However, as discussed in Part II, aiding and abetting requires knowledge and hence does not cover all the scenarios that command responsibility does.
[65] Of course, the commander could be liable in relation to a single crime if she ‘failed to prevent’ the crime as the contribution requirement would then be met.
[66] Part VII discusses this possibility.
[67] In such a case, there would be the requisite omission, the requisite state of mind, and a contribution to the crime, rendering her an accessory in a manner somewhat akin to aiding and abetting by omission: see Part III.
[68] A variation on this scenario is what we may call the ‘problem of the initial crime’. Assume that following her dereliction, further crimes do indeed occur. The commander may be properly liable for the subsequent crimes because her failure to punish prior crimes facilitates, encourages or has an effect on the subsequent crimes. However, she should not be liable for the initial crime or crimes (the crimes prior to the point where she knew or had reason to know that crimes were occurring) because she made no culpable contribution, by act or omission, to those crimes. As noted, this scenario assumes she took adequate preventive steps; if she did not, she could of course be held liable for the initial crimes, facilitated by her failure to meet her duty to prevent.
[69] Examples of such arguments in the context of command responsibility are discussed in the next section. Further examples are discussed in Robinson, ‘The Identity Crisis of International Criminal Law’, above n 30, 930–1.
[70] I say ‘did not make and could not have made’ a contribution because the contribution requirement plausibly only requires risk aggravation: see Part VII. A non-contributory failure to punish is still serious, and can and should be taken up by national legal systems, whose mandate is not restricted to the most serious international crimes. National systems are free to adopt the ‘separate offence’ approach and thus prosecute non-contributory derelictions as derelictions.
[71] See, eg, Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) in particular [21]–[22] (Judge Hunt), [23], [25], [38] (Judge Shahabuddeen). See also Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) in particular [19]–[21], [30]–[31] (Judge Liu), [8], [18]–[19] (Judge Schomberg).
[72] Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), opened for signature 8 June 1977, 1125 UNTS 3 (entered into force 7 December 1978) art 87(3).
[73] Robinson, ‘The Identity Crisis of International Criminal Law’, above n 30, 946–55.
[74] As Part VI will discuss, a lawmaker may certainly create a stand-alone offence criminalising such derelictions (with or without causal contribution to international crimes). It is doubted however whether the option is available to the international criminal courts and tribunals, given their applicable law and their practice of convicting persons as party to the underlying offence.
[75] A similar conclusion is reached in Antonio Cassese, International Criminal Law (Oxford University Press, 2nd ed, 2008) 243.
[76] See the judgments of Judges Shahabuddeen and Hunt in Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003); Prosecutor v Hadžihasanović (ICTY, Trial Chamber, Case No IT-01-47-T, 15 March 2006) [65].
[77] See, eg, Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) [53]–[56], and the partial dissent of Judge Shahabuddeen at [39]–[41].
[78] Ibid [37]–[56]. See also Theodor Meron, ‘Revival of Customary Humanitarian Law’ (2005) 99 American Journal of International Law 817, 824–6. While the approach does not directly reference the culpability principle, it does reflect concern for the legality principle.
[79] Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) [15]–[41] (Judge Shahabuddeen), [6]–[34] (Judge Hunt).
[80] See, eg, Joakim Dungel and Shannon Ghadiri, ‘The Temporal Scope of Command Responsibility Revisited: Why Commanders Have a Duty to Prevent Crimes Committed after the Cessation of Effective Control (2010) 17 University of California Davis Journal of International Law & Policy 1; Carol Fox, ‘Closing a Loophole in Accountability for War Crimes: Successor Commanders’ Duty to Punish Known Past Offences’ (2004) 55 Case Western Reserve Law Review 443; Christopher Greenwood, ‘Command Responsibility and the Hadžihasanović Decision’ (2004) 2 Journal of International Criminal Justice 598; Bing Bing Jia, ‘The Doctrine of Command Responsibility Revisited’ (2004) 3 Chinese Journal of International Law 1; Mettraux, above n 20, 190–2; Volker Nerlich, ‘Superior Responsibility under Article 28 ICC Statute: For What Exactly is the Superior Held Responsible?’ (2007) 5 Journal of International Criminal Justice 665; Chantal Meloni, ‘Command Responsibility: Mode of Liability for the Crimes of Subordinates or Separate Offence of the Superior?’ (2007) 5 Journal of International Criminal Justice 619; Roberta Arnold and Otto Triffterer, ‘Article 28: Responsibility of Commanders and Other Superiors’ in Otto Triffterer (ed), Commentary on the Rome Statute of the International Criminal Court: Observers’ Notes, Article by Article (Verlag C H Beck, 2nd ed, 2008) 795; David Akerson and Natalie Knowlton, ‘President Obama and the International Criminal Law of Successor Liability’ (2009) 37 Denver Journal of International Law and Policy 615, 627; Sepinwall, above n 5; Elies van Sliedregt, ‘Article 28 of the ICC Statute: Mode of Liability and/or Separate Offence?’ (2009) 12 New Criminal Law Review 420; Bakone Justice Moloto, ‘Command Responsibility in International Criminal Tribunals’ (2009) 3 Publicist 12; Stefan Trechsel, ‘Command Responsibility as a Separate Offence’ (2009) 3 Publicist 26; Barrie Sander, ‘Unravelling the Confusion concerning Successor Superior Responsibility in the ICTY Jurisprudence’ (2010) 23 Leiden Journal of International Law 105; Robert Cryer, ‘The Ad Hoc Tribunals and the Law of Command Responsibility: A Quiet Earthquake’ in Shane Darcy and Joseph Powderly (eds), Judicial Creativity at the International Criminal Tribunals (Oxford University Press, 2010) 159.
[81] See, eg, Fox, above n 80, 423 (‘loophole’); Mettraux, above n 20, 141 (‘gaping hole’, and citing Judge Hunt with approval); Akerson and Knowlton, above n 80, 645 (‘wrongdoers will escape reproach’). See also Dungel and Ghadiri, above n 80; Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) (Judges Shahabuddeen, Liu and Schomburg), as well as further examples below in Part VI.
[82] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [335]:
it should be immaterial whether he or she had assumed control over the relevant subordinates prior to their committing the crime. Since the Appeals Chamber, however, has taken a different view for reasons which will not be questioned here, the Trial Chamber finds itself bound …
See also Prosecutor v Ndindiliyimana (Judgment and Sentence) (ICTR, Trial Chamber II, Case No ICTR-00-56-T, 17 May 2011) [1960]–[1963] (offering a ‘sharp indictment of the jurisprudence’ which is ‘likely to allow superiors to escape criminal sanction for the role of their subordinates’) (‘Ndindiliyimana’); Prosecutor v Halilović (Judgement) (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [53] (‘Halilović’).
[83] Prosecutor v Sesay (Judgment) (Special Court for Sierra Leone, Trial Chamber I, Case No SCSL-04-15-T, 2 March 2009) [306] (‘RUF Case’):
this Chamber is satisfied that the principle of superior responsibility as it exists in customary international law does include the situation in which a Commander can be held liable for a failure to punish subordinates for a crime that occurred before he assumed effective control.
But see Prosecutor v Brima (Judgment) (Special Court for Sierra Leone, Trial Chamber II, Case No SCLC-04-16-T, 20 June 2007) [799] (‘AFRC Case’): ‘there is no support in customary international law for the proposition that a commander can be held responsible for crimes committed by a subordinate prior to the commander’s assumption of command over that subordinate’; Prosecutor v Fofana (Judgment) (Special Court for Sierra Leone, Trial Chamber I, Case No SCSL-04-14-T, 2 August 2007) [240] (‘CDF Case’): ‘The Chamber further endorses the finding of the ICTY Appeals Chamber that an Accused could not be held liable under Article 6(3) of the Statute for crimes committed by a subordinate before the said Accused assumed command over that subordinate’.
[84] Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008).
[85] Ibid [5], [8] (Judge Liu), [2] (Judge Schomberg).
[86] Ibid [3] (Judge Shahabuddeen). See also at [12].
[87] Ibid [167]:
The Appeals Chamber, Judge Liu and Judge Schomburg dissenting, declines to address the ratio decidendi of the Hadžihasanović Appeal Decision on Jurisdiction, which, in light of the conclusion in the previous paragraph, could not have an impact on the outcome of the present case.
[88] Ibid [19] (Judge Shahabuddeen). The passage is arguably ambiguous because Judge Shahabuddeen appears to contemplate only the mode of liability of commission and not other modes. See ibid [18]–[22]. See below Part VI.
[89] See, eg, ibid [18] (Judge Shahabuddeen).
[90] The ambiguities of the jurisprudence are discussed below in Part VI.
[92] Ibid 3–12, 74–82. The proposal that the general justifying aim may be entirely utilitarian has been questioned by other scholars. For example John Gardner notes that retributive considerations (inflicting appropriate suffering on wrongdoers for their wrongs) may be not only a constraint on punishment but rather a part of the aim and indeed the essence of punishment: see John Gardner’s ‘Introduction’ to Hart’s Punishment and Responsibility: Hart, above n 28, xii–xxxi. The point for now is that even if the justification for the system is consequentialist, deontological considerations at least constrain the pursuit of those aims.
[94] Immanuel Kant, Groundwork of the Metaphysics of Morals (Mary Gregor trans, Cambridge University Press, 1998) 38 [trans of: Grundlegung zur Metaphysik der Sitten (first published 1785)].
[95] See also Damaška, above n 16, 456.
[96] Prosecutor v Norman (Decision on Preliminary Motion Based on Lack of Jurisdiction (Child Recruitment)) (Special Court for Sierra Leone, Appeals Chamber, Case No
SCSL-2004-14-AR72(E), 31 May 2004) [14] (Judge Robertson).
[97] Lon L Fuller argues that the concept of persons as responsible agents is inherent in the enterprise of law, so that ‘[e]very departure from the principles of law’s inner morality is an affront to man’s dignity as a responsible agent’: Lon L Fuller, The Morality of Law (Yale University Press, revised ed, 1969) 162.
[98] Mark A Drumbl, Atrocity, Punishment and International Law (Cambridge University Press, 2007) 5–9, 38–9; Mark Osiel, ‘The Banality of Good: Aligning Incentives against Mass Atrocity’ (2005) 105 Columbia Law Review 1751, 1753, 1768.
[99] Robinson, ‘The Identity Crisis of International Criminal Law’, above n 30, 932–3, 962–3.
[100] See generally Robinson, ‘A Cosmopolitan Liberal Account of International Criminal Law’, above n 29.
[101] For example, alternative law-creation mechanisms and the absence of a legislature calls for reflection on the meaning of fair warning and the nature of the legality principle. Situations of mass criminality raise questions about deviance and choice: ibid.
[102] See Part I on the usage of the term ‘doctrinal’.
[103] Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998).
[104] Ibid [345], [396].
[105] Ibid [398].
[106] Ibid.
[107] The exception that the Trial Chamber noted was the work of Cherif Bassiouni, arguing that causal contribution was an essential element: see M Cherif Bassiouni and Peter Manikas, The Law of the International Criminal Tribunal for the Former Yugoslavia (Transnational, 1996) 350; M Cherif Bassiouni, Crimes against Humanity in International Criminal Law (Kluwer Law International, 1999) 372, cited in ibid.
[108] Blaškić (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [73]–[77], with similar issues also arising at [78]–[85].
[109] Ibid [76]–[77].
[110] Ibid [80]–[82].
[111] See, eg, Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [76]–[78]. See also Prosecutor v Brđanin (Judgement) (ICTY, Trial Chamber II, Case No IT-99-36-T, 1 September 2004) [280].
[112] I say ‘distinguish broadly’ because in a system that has fully internalised liberal values, deontological considerations would be part of contextual interpretation, informing and shaping formalistic doctrinal analysis, thus reducing the distinction that I draw here.
[113] One could of course rely on past cases if those cases expressly engaged with and addressed the issue of compliance with the fundamental principle in question and thus settled that particular issue. The point here is that past authority does not answer a principled challenge if the past authority was not engaged with the fundamental principle in question.
[114] Otto Triffterer, ‘Causality, a Separate Element of the Doctrine of Superior Responsibility as Expressed in Article 28 Rome Statute?’ (2002) 15 Leiden Journal of International Law 179; Mettraux, above n 20, 82–6, 236; Cassese, above n 75, 236–42; Greenwood, above n 80; Bantekas, ‘On Stretching the Boundaries of Responsible Command’, above n 34, 1208; Nerlich, above n 80, 672–3.
[115] For details on the Toyoda trial, see Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998) [339] (emphasis added).
[116] Ibid [341] (emphasis added).
[117] Ibid [396].
[118] Ibid [400]; Blaškić (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [76].
[119] Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998). The prosecution similarly rejected the possibility of causal nexus ‘as a matter of logic’: at [397]. Similarly, in Blaškić the Appeals Chamber noted that ‘disciplinary and penal action can only be initiated after a violation is discovered’ and found the defence argument that a contribution to crimes must be shown even under the failure to punish branch to be ‘illogical’: Blaškić (ICTY, Appeals Chamber, Case No IT-95-14-A, 29 July 2004) [83] (emphasis original).
[120] Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998) [396].
[121] For this form of argument, all in the context of successor commander liability, see, eg, Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [335]; Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) [7] (Judge Liu), [8] (Judge Schomberg); Ndindiliyimana (ICTR, Trial Chamber II, Case No ICTR-00-56-T, 17 May 2011) [1961].
[122] Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005).
[123] Ibid [78] describes command responsibility as a ‘sui generis form of liability’, which is distinct from the other modes of individual responsibility in that it ‘does not require a causal link’. See also Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [338].
[124] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [338].
[125] See discussion in Part II. There is also the important point that command responsibility may be satisfied by omission; however, as was shown in Part II, other modes of liability (such as aiding and abetting) can also be satisfied by omission by a person under a duty to act.
[126] Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [79].
[127] On omissions and causation, see below Part VII.
[128] Tadić (ICTY, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) [186].
[129] One might be able to uphold the no-contribution approach by reconceptualising the principle of culpability, but this would require careful deontological justification (see Part VII), not technical doctrinal arguments.
[130] Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) [32] (Judge Shahabuddeen).
[131] Ibid.
[132] For examples of jurisprudence, see: Fox, above n 80; Akerson and Knowlton, above n 80; Mettraux, above n 20; Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008); Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [335]; Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [53]; RUF Case (Special Court for Sierra Leone, Trial Chamber I, Case No SCSL-04-15-T, 2 March 2009) [306]; AFRC Case (Special Court for Sierra Leone, Trial Chamber II, Case No SCSL-04-16-T, 20 June 2007) [799]; CDF Case (Special Court for Sierra Leone, Trial Chamber I, Case No SCSL-04-14-T, 2 August 2007) [240].
[133] See numerous examples cited at above n 80.
[134] Kai Ambos, ‘Superior Responsibility’ in Antonio Cassese, Paola Gaeta and John R W D Jones (eds), The Rome Statute of the International Criminal Court: A Commentary (Oxford University Press, 2002) vol 1, 823, 850–2; Meloni, above n 80, 637.
[135] One would have to argue that the ‘object and purpose’ of the provision include complying with liberal principles, and/or that the ‘context’ includes the fact that the document is a liberal criminal justice instrument, both of which are plausible arguments.
[136] Similarly, Sander, above n 80, concludes that if command responsibility is a mode of liability, causation is needed; if it is a separate offence, causation of other crimes is not needed, because the dereliction itself can be the complete crime.
[137] The first option appears to be the model contemplated by jurists such as Judge Shahabuddeen. The second option appears to be the model contemplated by scholars such as Kai Ambos, since he regards it as a separate offence and yet would require causal contribution, along lines similar to those outlined in this paper: Ambos, ‘Superior Responsibility’, above n 134. The section option also matches the approach in legislation such as the Canadian war crimes legislation, which creates a separate offence of ‘breach of command responsibility’ and requires causal contribution to core crimes: Crimes against Humanity and War Crimes Act, SC 2000, c 24, ss 5, 7.
[138] Gesetz zur Einführung des Völkerstrafgesetzbuches [Act to Introduce the Code of Crimes against International Law] (Germany) 26 June 2002, Bundesgesetzblatt Jahrgang II, 2002 Nr 42, 2254, art 1 s 4(1): intentional omission to prevent the commission of a crime is a mode of liability; at art 1 ss 13, 14: failure to properly supervise and/or report crimes are separate offences. The German legislation recognises a mode of liability only where there is both a contribution and subjective mens rea; the latter requirement is plausibly cautious to ensure constitutionality. It was suggested above that the departure from subjective mens rea in ICL may be justifiable in a liberal system; however, the argument remains to be fully developed and judicially tested. Similarly, the Canadian implementing legislation, of which this author was one of the drafters, departed from the ‘mode’ approach in international law and adopted a ‘distinct offence’ approach, because the drafters could not be completely certain that the ‘should have known’ mental element would withstand scrutiny under Canadian constitutional principles: Kimberly Prost and Darryl Robinson, ‘Canada’ in Claus Kreß et al (eds), The Rome Statute and Domestic Legal Orders, Volume II: Constitutional Issues, Cooperation and Enforcement (Nomos, 2005) 52, 54–5.
[139] ICTY Statute art 7(1).
[140] Joshua Dressler, Understanding Criminal Law (LexisNexis, 5th ed, 2009) 381.
[141] Criminal codes can be divided into a ‘special part’, containing definitions of crimes and specific related rules, and a ‘general part’, laying out principles of general application, such as modes of liability.
[142] I use the term ‘adjunct crime’ rather than the more common term ‘inchoate crime’, because in inchoate (ie, ‘incomplete’) crimes the referent crime is not completed (hence there is only an ‘attempt’ at genocide or ‘incitement’ of genocide rather than principal or accessorial liability). If I used the concept of ‘inchoate’ crime, then an obvious counter-argument would be that command responsibility is not inchoate because the referent crime typically is completed. Thus, the broader term ‘adjunct crime’ gives the separate offence approach its strongest possible footing.
[143] In national systems, attempt and possibly incitement would be understood as adjunct offences (more specifically as inchoate offences) and this seems, subject to further reflection, a most plausible characterisation. See, eg, Dressler, above n 140, 379–419; Simester and Sullivan, above n 14, 269–78, 305–22. The wording of art 25 of the Rome Statute arguably presents them as a mode of liability; if that conception is to prevail then new thinking is needed about how to conceptualise them as modes of liability. See, eg, Albin Eser, ‘Individual Criminal Responsibility in Antonio Cassese, Paola Gaeta and John R W D Jones (eds), The Rome Statute of the International Criminal Court: A Commentary (Oxford University Press, 2002) vol 1, 767, 771, 817.
[144] Rome Statute art 28(a). See also at art 28(b).
[145] Article 7(1) contains only modes of liability properly so called, and no adjunct offences. Adjunct (inchoate) crimes such as attempt, conspiracy and incitement are listed only in the definition of crimes, specifically attached to the crime of genocide: ICTY Statute art 4(3). See also ICTR Statute art 2(3).
[146] See, eg, Sepinwall, above n 5, 265–9; Sander, above n 80; Meloni, above n 80; Cryer, above n 80.
[147] For examples referred to in Tribunal jurisprudence, see: Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998), citing French law (‘accomplices’): at [336]; citing Chinese law (‘accomplices’): at [337]; citing Yugoslav law (‘participant’): at [341]; citing United States v List (Judgment) (United States Military Tribunal, Nuremberg, Case No 7, 19 February 1948) (‘Hostage Case’) (‘held responsible for the acts of his subordinate commanders’): at [338].
[148] Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [42]–[53]. For discussion see: Sander, above n 80; Sepinwall, above n 5, 265–9, noting that the authorities invoked in Halilović give ‘overwhelming support’ to the mode of liability approach.
[149] Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [43], acknowledging that ‘[n]ational legislation enacted in the post World War II period, for example in Canada, France and Britain, considered command responsibility as a form of accomplice liability’ because it ‘amounted to encouragement or assistance of the subordinates in the commission of the crime’.
[150] Ibid [53].
[151] See, eg, Greenwood, above n 80, 603–4 (punished for the act); Sepinwall, above n 5, 267 (doctrinal history gives ‘overwhelming support for the mode of liability view’); Cryer, above n 80, 171–82 (form of liability for the underlying offence); Bantekas, ‘The Contemporary Law of Superior Responsibility’, above n 35, 577 (imputed liability); Cassese, above n 75, 206; Triffterer, above n 114, 229 (mode of participation); Arnold and Triffterer, above n 80, 843; Nersessian, above n 16, 89; Meloni, above n 80, 621–5; Darcy, above n 80, 391; W J Fenrick, ‘Some International Law Problems related to Prosecutions before the International Criminal Tribunal for the Former Yugoslavia’ (1995) 6 Duke Journal of Comparative and International Law 103, 111–2 (party to offence, not a separate offence); William H Parks, ‘Command Responsibility for War Crimes’ (1973) 62 Military Law Review 1; Yuval Shany and Keren R Michaeli, ‘The Case against Ariel Sharon: Revisiting the Doctrine of Command Responsibility’ (2002) 34 New York University Journal of International Law and Politics 797, 803, 829–31; Alexander Zahar, ‘Command Responsibility of Civilian Superiors for Genocide’ (2001) 14 Leiden Journal of International Law 591, 596 (mode of participation, not a crime); Michael L Smidt, ‘Yamashita, Medina, and Beyond: Command Responsibility in Contemporary Military Operations’ (2000) 164 Military Law Review 155, 168–9; Kevin Jon Heller, The Nuremberg Military Tribunals and the Origins of International Criminal Law (Oxford University Press, 2011) 271. But see Jia, above n 80, 34; Trechsel, above n 80; Ambos, ‘Superior Responsibility’, above n 134 (separate offence).
[152] Sander, above n 80, 122; Cryer, above n 80, 182.
[153] One might argue that the ICTY or ICC should treat contributory derelictions as a mode of liability (command responsibility) while also recognising a separate crime for non-contributory derelictions. The problem with that approach for the ICC is that the Rome Statute features a closed list of crimes and does not provide latitude to discover new crimes within it. For the ICTY, one might argue that the non-contributory derelictions might be read in as a war crime under art 3. This solution would not be tenable however, as war crimes occur only in armed conflict, whereas command responsibility also applies to genocide and crimes against humanity, which can occur outside armed conflicts.
[154] One example of such a passage is discussed in this section: Additional examples are discussed in Part VI(D).
[155] Robert Cryer et al, An Introduction to International Criminal Law (Cambridge University Press, 2nd ed, 2010) 398. See also Cryer, above n 80, 177–9 discussing the ‘entirely unreasoned’ passage.
[156] The passage is cited, for example, in Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) [19] (Judge Shahabuddeen).
[157] Prosecutor v Krnojelac (Judgement) (ICTY, Appeals Chamber, Case No IT-97-25-A, 17 September 2003) [171].
[158] Prosecutor v Krnojelac (Third Amended Indictment) (ICTY, Case No IT-27-95-I, 25 June 2001).
[159] See Part VI of the Krnojelac judgment: Prosecutor v Krnojelac (Judgement) (ICTY, Appeals Chamber, Case No IT-97-25-A, 17 September 2003) [108]–[498].
[160] Sander, above n 80, 116. In one decision, Orić, the Trial Chamber purported to convict the accused for a separate offence of ‘failing to discharge his duty to prevent’. The prosecution appealed on the grounds that this was a mischaracterisation of command responsibility, which is a mode of liability, and that the sentence failed to reflect its gravity as a mode of liability. The Appeals Chamber found that the factual findings for a command responsibility conviction had not been made and thus that the issue was moot: Orić Appeal (ICTY, Appeals Chamber, Case No IT-03-68-A, 3 July 2008) [79].
[161] See, eg, Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [78]: ‘the nature of command responsibility itself, as a sui generis form of liability, which is distinct from the modes of individual responsibility set out in Article 7(1), does not require a causal link’.
[162] See above Part II.
[163] It is conceivable that a deontological justification can be developed: see Part VII.
[164] Mettraux, above n 20, 37–47, 80–8. Mettraux rejects accessory liability as the appropriate category, inter alia, on the grounds that accessory liability requires knowledge. However, it is doubtful that the knowledge requirement is a fundamental defining feature of accessory liability. The theoretical basis for such a restriction has not been shown, and many national systems feature accessory liability with forms of fault other than knowledge.
[165] Including inchoate offences.
[166] I do not reject the possibility that there might be some hitherto undetected gap in the categories known to criminal theory. Indeed international criminal law may present new problems that help us discover those gaps: see Robinson, ‘A Cosmopolitan Liberal Account of International Criminal Law’, above n 29. I simply ask for greater precision as to what the alleged gap is.
[167] See, eg, Čelebići (ICTY, Trial Chamber, Case No IT-96-21-T, 16 November 1998) [333]: that commanders are ‘held criminally responsible for the unlawful conduct of their subordinates’.
[168] Later cases, such as Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [292]–[293], argue that ‘for’ the acts means ‘because of’ the acts. Likewise, Judge Hunt in Hadžihasanović explains that it is ‘not a direct responsibility for the acts of subordinate’ but rather a responsibility for the commander’s own acts or omissions: Hadžihasanović (ICTY, Appeals Chamber, Case No IT-01-47-AR72, 16 July 2003) [9]. See also Prosecutor v Aleksovski (Judgement) (ICTY, Trial Chamber, Case No IT-95-14/1-T, 25 June 1999) [72]; Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [54]. These passages are all perfectly defensible insofar as they are attempting to affirm that the commander is not deemed to be a perpetrator and that command responsibility is not vicarious liability.
[169] See, eg, Prosecutor v Aleksovski (Judgement) (ICTY, Trial Chamber, Case No IT-95-14/1-T, 25 June 1999) [72]: ‘superior responsibility … must not be seen as responsibility for the act of another person’ (emphasis added). Yet, at [67], the Trial Chamber stated that ‘[a] superior is held responsible for the acts of his subordinates if he did not prevent the perpetration of the crimes of his subordinates or punish them for the crimes’ (emphasis added). Similarly, Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [54] emphasises that the commander is not held liable ‘for’ the crimes but ‘because of’ the crimes. However, at [95], the Trial Chamber asserted that failure to punish is ‘so grave that international law imposes upon him responsibility for those crimes’ (emphasis added). This confusion arises from the vague uses of the term ‘responsible for the acts’. Indeed, the commander is not responsible as if the acts were her own conduct, but she can be responsible as an accessory.
[170] Consider for example Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [54] (citations omitted):
Thus ‘for the acts of his subordinates’ as generally referred to in the jurisprudence of the Tribunal does not mean that the commander shares the same responsibility as the subordinates who committed the crimes, but rather that because of the crimes committed by his subordinates, the commander should bear responsibility for his failure to act … a commander is responsible not as though he had committed the crime himself, but his responsibility is considered in proportion to the gravity of the offences committed.
What this paragraph attempts to describe is admirably captured in an existing legal category: accessory liability.
[171] Insofar as the passages are interpreted to mean that the commander is not charged as party to the crimes, then they would be an empirically untrue description of ICTY charging and conviction practice, as was discussed above.
[172] Stewart, above n 36, 25.
[173] For sophisticated examples of works that draw distinctions between different forms of command responsibility, see, eg, Meloni, above n 80; Nerlich, above n 80. Both plausibly distinguish between contributory and non-contributory derelictions and between those with and without subjective knowledge. These approaches are an advance over other approaches, because they grapple with culpability and acknowledge significant distinctions. My suggestion however is that simpler solutions can be found. Nerlich’s solution does not refer expressly to a separate offence, but would distinguish between holding the commander responsible for the crime and holding the commander responsible for the consequences of the crime: see at 680–2. The distinction is arguably a rather fine one, for example, liability for someone being murdered but not the murder, or for someone being raped but not the rape.
[174] See above, Part VI(B).
[175] I accept that interpreters can read counter-textual terms into a provision where it is necessary or valuable to do so and does not contradict fundamental principles, such as legality. Here however there is no need for a counter-textual reading.
[176] Risk aggravation is discussed further in Part VII.
[177] See Robinson, ‘A Cosmopolitan Liberal Account of International Criminal Law’, above n 29.
[178] Husak observes that there is no objective ‘fact’ that requires causation for culpability: Douglas N Husak, ‘Omissions, Causation and Liability’ (1980) 30 Philosophical Quarterly 318, 323. Hart and Honoré observe that ‘there is nothing to compel a legal system to accept a causal connection … as either necessary or sufficient for liability’, yet most systems generally do: see H L A Hart and Tony Honoré, Causation in the Law (Oxford University Press, 2nd ed, 1985) 63, 132.
[179] Hart and Honoré, above n 178, 62–8 (mentioning the exception of vicarious liability).
[180] Michael S Moore, Causation and Responsibility: An Essay in Law, Morals, and Metaphysics (Oxford University Press, 2009) 3–5, 110.
[181] Dressler, above n 37, 103.
[182] Gardner, above n 32, 127.
[183] Jerome Hall, General Principles of the Criminal Law (Bobbs-Merrill, 2nd ed, 1960) 248.
[184] Kadish, above n 21; Gardner, above n 32; Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37; Dressler, above n 37; Ashworth, above n 14, 415.
[185] Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 401.
[186] For example, Dressler argues that ‘causation’ is not required for accessory liability, but by ‘causation’ he refers to ‘but for’ causation; he acknowledges the need for assistance or influence: Dressler, above n 37, 102, 139. See also discussion by Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 402–7.
[187] Kayishema (ICTR, Trial Chamber II, Case No ICTR-95-1, 21 May 1999) [199].
[188] Orić (ICTY, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) [282].
[189] Ibid [284].
[190] Ibid [282].
[191] See, eg, Kadish, above n 21, 357; Mettraux, above n 20: ‘significant contributing factor’; Kai Ambos, ‘Article 28: Individual Criminal Responsibility’ in Otto Triffterer (ed), Commentary on the Rome Statute of the International Criminal Court: Observers’ Notes, Article by Article (Verlag C H Beck, 2nd ed, 2008) 743, 756–8. See also Gardner, above n 32, 137: ‘I am complicit only because my assistance actually assists [or] my encouragement actually encourages’.
[192] One of the strongest cases for this view is advanced by Moore who contends that an omission is a nothingness, or an absence, and an absence cannot produce effects; ‘nothing comes from nothing’: Moore, Causation and Responsibility, above n 180, 444–6. While Moore concludes that counterfactual dependency does not warrant the label ‘causation’, he holds that counterfactual dependency can give rise to liability. In this respect he reaches a similar endpoint to other scholars, albeit with significantly different labels: at 139–42, 351–4.
[193] See, eg, Fletcher, Basic Concepts of Criminal Law, above n 28, 67–9; George P Fletcher, Rethinking Criminal Law (Oxford University Press, revised ed, 2000) 585–625; Ashworth, above n 14, 418–20; Husak, above n 178, 160–5; Hart and Honoré, above n 178, 40, 447–9; Carolina Sartoria, ‘Causation and Responsibility by Michael Moore’ (2010) 119 Mind 475; Jonathan Schaffer, ‘Contrastive Causation in the Law’ (2010) 16 Legal Theory 259.
[194] And indeed those factors were jointly sufficient to produce the result, if we adopt a ‘causal efficacy’ approach that looks only at the physical forces that produced the result (naturalistic approach) and ignore the legal responsibilities sometimes placed upon humans to act and intervene in certain circumstances (normative approach).
[195] The debate arises because there are at least two major conceptions underlying causation. One conception looks at counterfactual dependence (the ‘but for’ test), examining what would have happened in an alternative universe without the variable in question. Another looks at the chain of events as they actually occurred, looking at the ‘causal energy’ or ‘causal efficacy’ of the forces sufficient to bring about the result. But causation is more subtle than either of these conceptions on its own. For example, it is well recognised that exclusive reliance on the counterfactual (‘but for’) test can at times generate absurd results. In ‘over-determined’ events, where there are multiple concurring sufficient causes, the ‘but for’ test would absurdly absolve all contributors, as each can accurately say that the event would have happened anyway. Thus, the ‘but for’ test cannot be the entirety of the test and we must resort to other tools. See, eg, Dressler, above n 37, 99–102; Hart and Honoré, above n 178, 122–5. Conversely, concerns about omissions tend to arise when one relies exclusively on concepts such as ‘causal energy’ or ‘causal efficacy’ and sets aside counterfactual analysis. This can also generate counterintuitive results, such as not conceding that failures by humans to fulfil their duties can have consequences, and that the omission to land the plane contributed to the crash, which arguably runs against common notions of causation and responsibility.
[196] See also Ambos, ‘Superior Responsibility’, above n 134, 860.
[197] Another justification for the more elastic approaches, which for reasons of space we will not explore here, is the problem of ‘moral luck’. It is often argued, with some plausibility, that persons should be held responsible for prohibited acts of risk-creation, and not for the lucky or unlucky fortuities of what consequences actually flow. This is a highly contested point in criminal law theory.
[198] See, eg, Hart and Honoré, above n 178, 51–9; Kadish, above n 21, 329–36; Dressler, above n 37, 127–28; Ashworth, above n 14, 415. An extreme form of belief in free will rejects the possibility of any causal contribution to the informed and voluntary decision of another human, but this position has fallen out of favour and it is well accepted that one may speak of humans influencing other humans: Stephen J Morse, ‘The Moral Metaphysics of Causation and Results’ (2000) 88 California Law Review 879; Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 414.
[199] Kadish, above n 21, 334; Hart and Honoré, above n 178, 51–7.
[200] Dressler, above n 37, 127–8; Christopher Kutz, ‘Causeless Complicity’ (2007) 1 Criminal Law and Philosophy 289, 294; Morse, above n 198, 883–6.
[201] This issue is sometimes resolved by defining the crime with increasing granularity; ie, the accomplice contributed to the crime in which the perpetrators entered by that particular entrance or used that particular weapon. Such solutions are not perfectly satisfactory: see Kutz, above n 200, 297; Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 406–7.
[202] Kadish, above n 21, 358.
[203] Ibid 395.
[204] Kutz, above n 200, 294 (‘might have made a difference’).
[205] One could debate whether it is a distinct vein or an alternative articulation of a similar intuition. For present purposes it does not matter whether the two approaches are different or effectively the same, or whether one is preferable to the other, because Tribunal jurisprudence on the ‘failure to punish’ branch falls foul of either test.
[206] Ambos, ‘Superior Responsibility’, above n 134, 860–1. For risk aggravation theory as appropriate in accessory liability, see Ambos, ‘Article 28: Individual Criminal Responsibility’, above n 191, 758–9.
[207] Roberta Arnold, ‘Command Responsibility: A Case Study of Alleged Violations of the Laws of War at Khiam Detention Centre’ (2002) 7 Journal of Conflict and Security Law 191, 207–8.
[208] Hart and Honoré, above n 178, lxii–lxv, 6, 81–2, 286, 478–95. Moore would decline to apply the label of ‘causation’ to risk-aggravation, but regards risk aggravation as a sufficient basis for liability, which again brings us to a similar endpoint, albeit with different terminology and conceptualisations: Moore, ‘Causing, Aiding, and the Superfluity of Accomplice Liability’, above n 37, 432–40. See also Moore, Causation and Responsibility, above n 180, 308–10.
[209] Hart and Honoré, above n 178, 59–61, 80–2.
[210] A potential alternative method would be to consider the template of regulatory enterprise liability, such as the liability of owners of enterprises for regulatory offences and of newspaper owners for libels. Such regimes apply to regulatory offences that are not regarded so as to apply a ‘stigma’ to be borne by the accused. The culpability principle requires a personal contribution for those crimes that bear a stigma, such that the person is being ‘blamed’ for the crime and exposed to significant punishment. Thus, this model does not seem viable as a solution in a liberal system of justice. As Francis Sayre has written:
Where the offense is in the nature of a true crime, that is, where it involves moral delinquency or is punishable by imprisonment or a serious penalty, it seems clear that the doctrine of respondeat superior must be repudiated as a foundation for criminal liability. For it is of the very essence of our deep-rooted notions of criminal liability that guilt be personal and individual; and in the last analysis the … sense of justice … is the only sure foundation of law.
Francis Bowes Sayre, ‘Criminal Responsibility for the Acts of Another’ (1930) 43 Harvard Law Review 689, 717. See also Timothy Wu and Yong-Sung Kang, ‘Criminal Liability for the Actions of Subordinates — The Doctrine of Command Responsibility and Its Analogues in United States Law’ (1997) 38 Harvard International Law Journal 272, 278–85.
[212] Ibid 304.
[213] Ibid. The reasons include, inter alia, that ratification expresses an agent’s identification with the acts, that it reveals the dangerousness of the ratifying accomplice, and that, in an institutional context, ratification may have future effects and exacerbate the dangerousness of future principals.
[214] Sepinwall, above n 5, 255.
[215] Ibid 296.
[216] Ibid 286–302.
[217] Mettraux, above n 20, 43, 80, 86.
[218] United States v von Leeb (United States Military Tribunal, Nuremberg, Case No 12, 28 October 1948).
[219] Ibid 568. The commander was Wilhelm von Kuechler.
[220] Prosecutor v Sawada (United States Military Commission, Shanghai, Case No 25, 15 April 1946).
[221] Halilović (ICTY, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) [95].
[222] Dubber, above n 21, 979–81, 997–8; Damaška, above n 16, 468–9.
[223] Simester and Sullivan, above n 14, 195 n 9.
[224] See, eg, American Law Institute, Model Penal Code (1962) § 242.3; Damaška, above n 16, 468–9; Dubber, above n 21, 998; Eser, above n 143, 806–7; Ambos, ‘Article 28: Individual Criminal Responsibility’, above n 191, 769–70; Fletcher, Rethinking Criminal Law, above n 193, 646.
[225] Dubber, above n 21, 998: the aider after the offence ‘bears no causal responsibility for the offence; his conduct does not even rise to the level of a contributory cause’. I should reiterate that steps taken before completion of the crime, such as agreeing in advance to help ensure that the perpetrators escape justice, may of course facilitate or encourage the crime, and thus can satisfy the causal contribution requirement. See Part III.
[226] ‘Draft Articles on Responsibility of States for Internationally Wrongful Acts’ in International Law Commission, Report of the International Law Commission, UN GAOR, 56th sess, Supp No 10, UN Doc A/56/10 (2001) 43; United States Diplomatic and Consular Staff in Tehran (United States of America v Iran) (Judgment) [1980] ICJ Rep 3, 34–5 [71]–[74].
[227] Sayre, above n 210, 708.
[228] Kadish, above n 21, 354.
[229] Hart and Honoré discuss cases suggesting ‘ratification’ of past criminal acts as ‘doubtful cases’ and express doubt that approval constitutes a separate ground of liability. They also note, however, that even if such a theory can be sustained, it would be an alternative to causation: Hart and Honoré, above n 178, 387–8.
[230] For a discussion in a different context, see, eg, George P Fletcher, ‘Collective Guilt and Collective Punishment’ (2004) 5 Theoretical Inquiries in Law 163, 169–72.
[231] Sepinwall, above n 5, 295, tracing the concept to Kant.
[232] Karl Jaspers, The Question of German Guilt (E B Ashton trans, Greenwood Press, 1978) 73–4 [trans of: Die Schuldfrage (first published 1947)].
[233] Ibid.
[234] Fletcher observes that to ground criminal liability on notions of pollution would arguably ‘jar our modern sensitivities’: Fletcher, ‘Collective Guilt and Collective Punishment’,
above n 230, 171.
[235] See, eg, Kutz, above n 200, 300–4; Dressler, above n 37, 109–11; Kadish, above n 21, 355: ‘we identify with those actions by intentionally helping or inducing [the perpetrators] to do those actions … in other words by extending our will to those actions’.
[237] An alternative would be to narrow the mental element of command responsibility in order to sustain a will-based theory, but this may be difficult to reconcile with doctrine and with the distinguishing feature of command responsibility (holding the commander liable for failing to intervene where she ‘had reason to know’ or ‘should have known’ of crimes). A second alternative would be to locate ‘will’ even in the more permissive mental elements of command responsibility. A third alternative would be to build a justification for liability for failures to punish that are either: (a) with subjective knowledge; or (b) make causal contributions, but the latter is complex and not easy to reconcile with the textual enactments.
[238] See Part III.
[239] See Part III.
[240] The scenario was discussed in more detail in Part III(D).
[241] Another example would concern the initial crimes in a series of crimes. Where there was no ‘failure to prevent’ and hence liability can only be based on ‘failure to punish’, the commander may be liable for crimes following her dereliction but should not be liable for the initial crimes because she did not contribute to them.
[242] See Part V(E).
[243] A counter-argument would be that the ‘should have known’ standard is inadequate to meet the subjective aspect of the culpability principle, and hence one could argue that a separate offence is still needed for criminal negligence situations. I believe however that a deontological case can be made for the mental element in art 28 of the Rome Statute as I have briefly touched upon in Part II(B).
[244] See, eg, Akerson and Knowlton, above n 80, 630 (‘obvious flaw’); Fox, above n 80, 480 (‘weaknesses and limitations’); Emily Langston, ‘The Superior Responsibility Doctrine in International Law’ (2004) 4 International Criminal Law Review 141, 161 (‘retreat’); Jordan Paust, ‘Content and Contours of Genocide, Crimes against Humanity and War Crimes’ in Siehno Yee and Wang Tieya (eds), International Law in the Post-Cold War World: Essays in Memory of Li Haopei (Routledge, 2001) 289, 305 (‘problem’).
[245] See Part V.
[246] See Part III.
[247] Bemba (ICC, Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009).
[248] See Part VI.
[249] Bemba (ICC, Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009) [426]. See also at [405]: ‘responsible for the prohibited conduct of his subordinates’ and at [427], emphasising that it is not a form of ‘strict liability’ but rather the ‘attribution of criminal responsibility’ for core crimes depends on the existence of the required state of mind or degree of fault.
[250] Ibid. The disposition follows at [501].
[251] The Chamber’s requirement that the commander’s omission ‘increased the risk of the commission of the crimes charged’ clearly implies that the commander is ‘charged’ with the crimes perpetrated by the subordinates: ibid [425]. Structurally, the decision separates the ‘crimes charged’ (Heading V) and the question of ‘individual criminal responsibility’ for those crimes (Heading VI), and discusses command responsibility under the latter.
[252] Ibid [418].
[253] Ibid [419].
[254] Ibid [420]–[423].
[255] Ibid [422]. Whereas the ICTY Statute refers to two branches, ‘to prevent’ or ‘to punish’, art 28 of the Rome Statute refers instead to three branches: to prevent, to repress or to submit the matter to competent authorities for investigation and prosecution: Rome Statute arts 28(a)(ii), 28(b)(iii). This difference was noted and discussed by the Chamber. The Chamber plausibly interpreted ‘to repress’ as referring to: (a) a duty to stop ongoing crimes; and (b) a duty to punish crimes herself where she has authority to do so. The first is, in effect, a more precise sub-category of the duty to prevent, and the second is a more precise sub-category of the duty to punish, with the Rome Statute expressly distinguishing the situation where the commander cannot punish the crime herself but must refer it to other authorities. This paper will continue using the two concepts of ‘prevent’ and ‘punish’ because the additional sub-categories acknowledged in the Rome Statute are granulations that do not materially alter the analysis here. The existence of the sub-categories is implicitly acknowledged.
[256] Ibid [425].
[257] See also discussion in Kai Ambos, ‘Critical Issues in the Bemba Confirmation Decision’ (2009) 22 Leiden Journal of International Law 715.
[258] Bemba (ICC, Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009) [425]:
contrary to the visible and material effect of a positive act, the effect of an omission cannot be empirically determined with certainty. In other words, it would not be practical to predict exactly what would have happened if a commander had fulfilled his obligation to prevent crimes.
[259] The view expressed in Bemba may overestimate the clarity of the impact of acts. The counterfactual analyses of acts and omissions equally involve imagining a hypothetical alternative world, so neither trades in certainties. It is true that the impact of an act may often seem clear; for example, stabbing a person causes a sudden deterioration in health which would not likely have happened otherwise at that moment. However, the impact of acts can equally be extremely difficult to assess; for example, did words of encouragement encourage the crime, did one blow among many other blows hasten the death. Conversely, the impact of omissions can be quite clear, as in the case of the pilot choosing to slump passively during a routine landing. It is suggested that the difficulties in isolating the impact of the commander’s omission may be rooted more in the difficulties of discerning impacts on human behaviour (see Part VII) than on any inherent difference between acts and omissions.
[260] Elsewhere I have criticised unduly simplistic and blinkered approaches to teleological interpretation which focus on the single aim of victim protection to the exclusion of all other aims: see Robinson, ‘The Identity Crisis of International Criminal Law’, above n 30. The teleological approach employed here acknowledges the multiplicity of aims, including the commitment to comply with fundamental principles.
[261] Rome Statute arts 21(1)(b), 21(1)(c) expressly allow reference to customary law and general principles of law. See also Vienna Convention on the Law of Treaties, opened for signature 23 May 1969, 1155 UNTS 331 (entered into force 27 January 1980) arts 31(2), 31(3).
[262] See Part VII. See also Osiel, above n 98, 1780–1; Nerlich, above n 80, 673.
[263] Bemba (ICC, Pre-Trial Chamber II, Case No 01/05-01/08, 15 June 2009) [424].
[264] Ibid.
[265] See ibid.
[266] Where liability is based on failures to punish, the wording of art 28 arguably requires two derelictions. The chapeau of art 28(a), the contribution requirement, requires that a failure to exercise control properly (in this case, a failure to punish) contributed to crimes. In addition, the text of art 28(a)(ii) arguably also requires that the commander fail to prosecute the crimes charged (the subsequent crimes) as well. The second failure is not required by the culpability principle: it ought to be sufficient that a failure to punish crime A contributed to crime B, thus grounding accessory liability for crime B. However the drafting of art 28(a)(ii) arguably requires, in addition, a failure to punish crime B to meet the textual requirements.
This possibility was foreshadowed in Professor Otto Triffterer’s early and careful dissection of art 28, which suggested that two derelictions may be required by the text: Triffterer, above n 114. That conclusion seems insightful and correct with respect to liability based on ‘failures to punish’. However, where liability is based on a ‘failure to prevent’, a single dereliction could satisfy both the requirement of a dereliction (the failure to prevent) and the contribution requirement. This conclusion is supported by the Bemba decision, which concluded that a ‘failure to exercise control properly’ means a failure to prevent or a failure to punish. The point is admirably explained in Nerlich, above n 80, 678.
[267] Bemba (ICC, Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009) [424].
[268] There could conceivably be circumstances where the derelictions of a commander had no possible impact on subordinate conduct, but in such circumstances it would seem that the commander would have so completely lost influence and control that the ‘effective control’ requirement would not be met.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/MelbJlIntLaw/2012/2.html