1. Introduction
As the International Criminal Court moves into its second decade of existence, questions involving the proper methodological approach to the interpretation of the Rome Statute have never been more pressing. Differences regarding the meaning of the ‘policy’ element in Article 7, the evidence required to constitute a ‘reasonable basis’ under Article 15, and the meaning of ‘commission’ in Article 25 have generated lively debates at the Court and amongst scholars and practitioners. The correct scope of Article 25, in particular, has led to divergent views at the Court and spawned a vast literature on modes of liability in international criminal law. This article explores the debate surrounding Article 25's meaning and proposes seven canons of ICC treaty construction that may serve as the basis of a principled interpretation of the substantive law of the Rome Statute. Building upon classic methodologies of treaty interpretation, these seven canons nonetheless incorporate the specificity of the Rome Statute regime, including the rules regarding sources of law in Article 21 and the legality principle in Article 22(2).
Why Article 25? First, it is particularly important for practical and expressive reasons. Its expressive function is to describe the actions of the accused that may subject him or her to potential criminal liability. In this way it may put not only the accused but also other perpetrators on notice of what behaviour is prohibited and allow the lay reader to understand what conduct the Court is pursuing. Its practical function is to permit the prosecutor, defence counsel, judges, and the accused to understand just what the individual indicted by the Court is said to have done, and to ensure that the proof adduced in a particular case establishes, beyond a reasonable doubt, the culpability of the accused. Second, like other provisions of the Statute which were negotiated and understood by individuals hailing from diverse legal traditions, Article 25 is like a legal Rorschach blot,Footnote 1 taking on a different meaning depending upon the underlying legal training, tradition, and even policy-orientation of those seeking to interpret it. Finally, there is now a well-developed – and divergent – jurisprudence from the Court on Article 25 and the issue of its interpretation is now squarely before the Appeals Chamber in the Lubanga case,Footnote 2 rendering this analysis particularly timely.
In Prosecutor v. Thomas Lubanga Dyilo,Footnote 3 the ICC's first completed trial and judgment, the majority relied upon a line of jurisprudence emanating from the Court's pre-trial chambers assigning a complex meaning to Article 25(3)(a), based upon Claus Roxin's ‘control of the crime’ theory (explored below in Section 3.3).Footnote 4 Judge Adrian Fulford took issue with the majority's analysis, arguing that its interpretation of Article 25(3)(a) was erroneous and that it had improperly imported the ‘control of the crime’ theory into Article 25.Footnote 5 Subsequently, in Prosecutor v. Mathieu Ngudjolo Chui,Footnote 6 Judge Christine Van den Wyngaert appended a concurring opinionFootnote 7 to the decision of Trial Chamber I acquitting Ngudjolo, aligning herself with Fulford's position, upon which she substantially elaborated. The recent judgment in Prosecutor v. Germain Katanga saw Van den Wyngaert once more reiterating her position in her dissent,Footnote 8 while the majority diverged in a significant way from prevailing interpretations of Article 25.Footnote 9
The Fulford and Van den Wyngaert opinions in Lubanga and Ngudjolo articulated the unease that many observers of the ICC's early case law had felt (including these authors) that the ICC had embarked upon a methodological approach to Article 25 that seemed to fit uneasily with the plain language of the Rome Statute, the work of the ad hoc international criminal tribunals, and customary international law. The continued debate in the Katanga judgment gives rise to the same concern. But given the differing views on Article 25 at the Court, it is timely to ask, what principles should guide the Court in its interpretative work?
This article proposes an answer, at least as regards provisions of the Statute which, like Article 25(3)(a), form part of the ‘substantive’ criminal law of the Rome Statute referred to in Article 22.Footnote 10 Rather than rehash the terms of what has become a highly contested and inconclusive debate, this article takes a different approach, using first principles to resolve the interpretative difficulty postulated by Article 25.
This article begins by outlining the rules and sources of law relevant to interpreting the Rome Statute and highlighting the inherent tensions between these. On the basis of this analysis, we propose seven canons of construction that may serve as the foundations of a principled interpretation of the Rome Statute. Section 3 charts the evolution of Article 25 through customary international law, the jurisprudence of the ad hoc tribunals, and the drafting of the Rome Statute, before exploring the emerging jurisprudential divisions at the ICC on this provision. Endeavouring to resolve these divisions, Section 4 applies the seven canons to Article 25(3) in particular. We conclude that it was improper for the ICC to incorporate the so-called ‘control theory’ into its interpretation of Article 25, regardless of the merits of that theory in describing the complexity of system criminality.Footnote 11 Indeed, doing so violates – to varying degrees – each one of the seven canons we suggest should guide the Court in its work. We also reject the notion that Article 25(3) sets out a well-defined hierarchy of criminal participation, and six of the seven canons we propose suggest that the sharp distinction between principals and accessories is neither required by the text of the Rome Statute nor customary international law. Instead, the plain meaning of Article 25(3) suggests neither any particular hierarchy implicit in its application, nor any importation of the principal/accessory distinction from municipal law. Rather, we suggest, sub-paragraphs (a) through (d) set out various overlapping forms of criminal participation, any one of which may give rise to criminal responsibility under the Statute, responsibility that will then be assessed in terms of culpability at the sentencing stage of the proceedings.
We recognize that treaty interpretation partakes ‘as much of art (the art of judgment) as of science (the science of law)’.Footnote 12 Our approach is therefore modest, drawing from well-established understandings of international treaty interpretation and the role of the judicial function. At the same time, we suggest guidelines that enable the ‘Rorschach blot’ of Article 25, capable of so many divergent interpretations, to become uniformly and consistently understood and interpreted, and note the utility of ‘identifying and applying different hermeneutics’ – including canons of construction – to different treaty regimes.Footnote 13 Moreover, it is important to note that this method does not forbid the Court from adopting a more teleological approach to elements of the Rome Statute not constrained by the application of Article 22(2), particularly the provisions of the Statute that touch upon the its nature as a constitutive document.Footnote 14 The difficulty and complexity of the Rome Statute is extraordinary, making it vital that the Court develop a consistent methodology of interpretation, particularly as regards the ‘criminal code’ embedded in the treaty. This will assist the Court's judiciary in being appropriately creative where the Statute leaves difficult and open-textured problems of interpretation, but help them to avoid charges of ‘judicial activism’ in going beyond interpretations that are foreseeable and predictable.Footnote 15 This may strengthen both the expressive value and common understanding of the Rome Statute and enhance its legitimacy in the eyes of the many constituencies it was established to serve.
2. The interpretative methodology of the Rome Statute
The prevailing jurisprudential interpretation of Article 25 has developed in an unconvincing manner, with scant attention paid to explicitly distilling and applying an interpretative methodology suited to the Rome Statute.Footnote 16 As the following analysis shows, traditional interpretative methodologies (including a straightforward application of the Vienna Convention on the Law of Treaties)Footnote 17 do not fit neatly with the unique characteristics of the Rome Statute.Footnote 18 After all, like the constitutive treaties of other international organizations, the Rome Statute takes the form of an international treaty, but many provisions have constitutional status, with others performing the function of legislation, in particular Articles 6 through 8 bis, which incorporate a criminal code within the text of the treaty itself.Footnote 19 Indeed, depending upon which provision is sought to be interpreted, a ‘plain meaning’, subjective, or teleological (or effective) approach may be appropriate.Footnote 20 The Rome Statute itself does not prescribe an interpretative methodology, although the thrust of Articles 21 and 22 suggest an emphasis upon textual interpretation as opposed to expansive interpretation of substantive law.Footnote 21 Moreover, the Statute includes provisions on applicable law and certain guides to the Court in interpreting the Statute, such as Article 21(3)'s admonition to apply and interpret Rome law consistently with international human rights norms. The following sections are devoted to identifying a methodological framework that can be used to interpret provisions of the Rome Statute subject to Article 22(2)'s requirement of strict construction, which we assume, as Judge van den Wyngaert did,Footnote 22 includes Article 25 in its ambit, although an argument can certainly be made that Article 25 is not part of the crime's definition and therefore not subject to Article 22(2)'s application.Footnote 23 This is then applied to the debate over the meaning of Article 25.
2.1. Article 21: The applicable law
Article 21 of the Rome Statute tasks the Court with applying first, the Statute itself, the Elements of Crimes and its Rules of Procedure and Evidence (Article 21(1)(a)); second, where appropriate, applicable treaties and the principles and rules of international law, including the established principles of the international law of armed conflict (Article 21(1)(b)); and third, ‘failing that’,
general principles of law derived by the Court from national laws of legal systems of the world including, as appropriate, the national laws of States that would normally exercise jurisdiction over the crime, provided that those principles are not inconsistent with [the] Statute and with international law and internationally recognized norms and standards. (Article 21(1)(c)).
Article 21(1)(a) is further qualified by Article 9 which identifies the Elements as a subsidiary source of law.Footnote 24 The Appeals Chamber has ruled that the application of the second and third subsidiary sources of law can only be justified where there is a lacuna in the text of the Statute.Footnote 25 There has been some debate over what constitutes the subsidiary sources referred to in paragraphs 1 (b) and (c);Footnote 26 however a general framework can be identified.
‘Applicable treaties’, referred to in sub-paragraph (1)(b), have been held by the Appeals ChamberFootnote 27 to include, inter alia, the Vienna Convention on the Law of Treaties (VCLT).Footnote 28 This has been criticized on the grounds that the interpretation of the Statute as an internal document of the Court is different from the interpretation of a treaty between states.Footnote 29 While a comprehensive treatment of this debate is beyond the scope of this article, it is worth noting that the VCLT is both widely recognized as a codification of customary international law and a persuasive and helpful rubric for the judges of the Court to apply.Footnote 30 The VCLT provides a more explicit interpretive framework than the Rome Statute itself, and for this reason reliance on it by judges and scholars seems both natural and logical. However, as discussed further in Section 2.2 below, there may be some tension between certain rules of the VCLT and their application to the criminal code within the Rome Statute.
It is generally understood that ‘principles and rules of international law’, as per paragraph 1(b), incorporates customary international law into the Rome Statute,Footnote 31 but as a gap filler, subsidiary to the text of the Statute itself. Thus, the possibility arises that earlier case law explicitly based upon customary international law could be contrary to the Rome Statute, which is undoubtedly why Judge Van den Wyngaert recently worried aloud that it may be impermissible for the ICC, unlike the ICTY Chambers, to ‘draw [directly] on customary international law in order to interpret modes of liability under their Statute’.Footnote 32 Yet, at the same time, the Statute must be understood within the context of the rules of international humanitarian law that it codifies. For example, to understand the meaning of the word ‘civilian’ in Article 8(2)(b)(i) or (ii) (prohibiting attacks against the civilian population and civilian objects), one must refer to authoritative commentaries on the laws of war – the dictionary would be of limited assistance. For this reason, as noted below, it is the view of these authors that, whenever possible, ICC judges should align their jurisprudence with customary international law to avoid the fragmentation of international criminal law and its decreased legitimacy, particularly as regards referrals of situations to the Court involving ICC non-states parties that have not accepted the Court's jurisdiction explicitly.Footnote 33
As a last resort the Court can turn to ‘general principles of law derived by the Court from national laws of legal systems of the world’.Footnote 34 William Schabas suggests that Article 21(1)(c) should be interpreted as ‘an invitation to consult comparative criminal law as a subsidiary source of norms’.Footnote 35 The task of identifying a ‘general principle of law’ is exceedingly difficult given the diversity of the world's legal systems and presents its own methodological challenges.Footnote 36 However, the Erdemovic case suggests some guidance as to how such an analysis may be accomplished:
[O]ur approach will necessarily not involve a direct comparison of the specific rules of each of the world's legal systems, but will instead involve a survey of those jurisdictions whose jurisprudence is, as a practical matter, accessible to us in an effort to discern a general trend, policy or principle underlying the concrete rules of that jurisdiction which comports with the object and purpose of the establishment of the International Tribunal.Footnote 37
The Court has generally been resistant to arguments based upon national practice as a source of general principles and some delegations expressed a view during the drafting process that, ‘as a matter of principle, no reference to any national law of states should be made. The Court ought to derive its principles from a general survey of legal systems and their respective laws’.Footnote 38
2.2. The application of the VCLT to the Rome Statute, Article 22 of the Rome Statute, and the principle of legality
The ‘kaleidescopic’ difficulties surrounding application of the Vienna Convention notwithstanding,Footnote 39 there is significant judicial support for recourse to the VCLT when interpreting the Rome Statute. Article 31(1) requires that ‘[a] treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in light of its object and purpose’. Article 32 provides that supplementary means of interpretation include the preparatory work of the treaty and the circumstances of its conclusion. Recourse to these tools is available when an interpretation according to Article 31 ‘leaves the meaning ambiguous or obscure’ or ‘leads to a result which is manifestly absurd or unreasonable’.Footnote 40 Although this suggests a two-step process, the International Law Commission recently proposed that ‘[t]he interpretation of a treaty consists of single combined operation, which places appropriate emphasis on the various means of interpretation indicated, respectively, in Articles 31 and 32’.Footnote 41
Relying solely on rules of interpretation derived from the VCLT, however, is inappropriate within the Rome Statute system as regards the substantive law of the Statute, as Article 22(2) explicitly incorporates the principle of legality or nullum crimen sine lege, providing:
The definition of a crime shall be strictly construed and shall not be extended by analogy. In case of ambiguity, the definition shall be interpreted in favour of the person being investigated, prosecuted or convicted.
Some commentators have noted that application of the principle of legality contradicts the rules of interpretation in the VCLT in a number of circumstances,Footnote 42 such as the reference to subsequent practice in Article 31(3) potentially conflicting with the rules on non-retroactivity.Footnote 43 Moreover, referencing the ‘object and purpose’ of the Rome Statute can lead to interpretations of provisions that seek to protect more victims but at the expense of the rights of the accused.Footnote 44
While it is true that the principle of legality, strictly applied, is often at odds with the teleological method of interpretation, the fact remains that ambiguities in the law must be resolved and this can only be done through judicial interpretation. Moreover, it is not possible (or appropriate) to ‘elevate strict construction over every other goal of the ICC Statute, including substantive justice’.Footnote 45 Rather, what is required by the principle of legality is that a judicial interpretation is ‘reasonably foreseeable’ and consistent with the essence of an offence.Footnote 46 Indeed, the art of interpretation is important because ‘it is impossible to foresee and point out all the particular cases that may arise’.Footnote 47 This is particularly the case with Article 25 given its relatively eclectic combination of national and international sources of law, as is further explained below. The challenge then is to balance these interpretative tensions and distill a methodological framework of interpretation that can be used to give meaning to the Rome Statute. How can the judges of the Court engage in much needed interpretative – and creative – work while avoiding the charge that they are engaged in improper judicial activism?Footnote 48
2.3. Seven canons of ICC interpretation
Because traditional methods of treaty interpretation may not always be suited to the Rome Statute, this section suggests seven core principles or canons of ICC interpretation, applicable to the criminal code within the Statute, that is, those provisions of the Statute that must be read in light of Article 22(2). These canons endeavour to balance the inherent tensions between the principle of legality, traditional methods of interpretation, and the need to resolve ambiguities in the Rome Statute, as well as consider the ICC's important and central role in law-making as the world's first permanent international criminal court. Additionally, although Article 21(1)(a) and (b) and (c) establish, by their terms, an interpretative hierarchy, it is important to view these provisions and the canons articulated here in a holistic fashion, as it may be difficult at each stage to have a complete understanding as to a provision's meaning and proper interpretation without doing so. Thus, analogous to Article 38(1) of the Statute of the International Court of Justice, they are listed ‘in the order that they would normally present themselves to the mind of an international judge’, and in this way form a practical methodology, as follows:Footnote 49
Canon 1: Fidelity to the text and reliance upon plain meaning including ordinary principles of treaty interpretation such as good faith and consideration of context should guide the Court's construction of a particular provision.
Canon 2: Provisions should be construed to be faithful to the object and purpose of the ICC Statute, consistent with the legality principle embodied in Article 22(2).
Canon 3: Where the meaning of a particular provision remains ambiguous or obscure or leads to a result that is manifestly absurd or unreasonable, the travaux préparatoires may be consulted.
Canon 4: If gaps remain in the interpretation of a particular provision, the Court should look to Article 21(1)(b) sources of law, including the jurisprudence of the ad hoc international criminal tribunals, and, failing that, Article 21(1)(c) sources.
Canon 5: All provisions should be construed with the objective of protecting the rights of the accused and ensuring that the application of the Statute is consistent with internationally recognised human rights.
Canon 6: The interpretation adopted should enhance judicial efficiency and the effectiveness of the ICC trial system, without compromising the values expressed in Canon 5.
Canon 7: The interpretation of a particular provision should enhance the expressive and normative function of international criminal law by rendering it transparent and comprehensible and reducing opportunities for fragmentation.
Each of these canons is discussed further below, and subsequently, applied to the interpretation of Article 25, in Section 4.
2.3.1. Canon 1: Fidelity to the text and reliance upon plain meaning including ordinary principles of treaty interpretation such as good faith and consideration of context should guide the Court's construction of a particular provision
Consistent with Article 21(1) of the Rome Statute, the first and overriding principle of ICC interpretation should be a fidelity to the text and reliance upon the plain meaning of the words of the Statute before moving to any subsidiary sources.
Determining the ‘plain meaning’ of a text, however, is often easier said than done. While resort to the dictionary is sometimes useful, and was a technique employed by the judges in Lubanga regarding the meaning of ‘enlistment’ and ‘conscription’,Footnote 50 this is a methodology to be sparingly employed given that a text like the Rome Statute is ‘authentic’ or official in six languages and is a highly complex instrument with ancillary texts like the Rules of Procedure and Evidence and Elements of Crimes that complete its meaning.Footnote 51 Rare indeed will be the utility of such an exercise, and it should not be used to disguise overly expansive or restrictive interpretations of provisions that are easily understood without resort to extrinsic aides. A judge is not required ‘to superimpose customary international law upon every treaty provision’ but the interpreter does need to be ‘sensitive to the fact that in treaty-making, no peculiar domestic construction of legal terms is contemplated or acceptable’.Footnote 52 In other words, the treaty should be read and understood against the backdrop of the international legal order from which it both emanates and to which it contributes.
This is particularly difficult at times with multilateral treaties and it may often be important to consult a provision's meaning in multiple languages,Footnote 53 as Judge Kaul did quite appropriately in the Kenya Article 15 decision determining the meaning of ‘State or organisational policy’ in Article 7(2)(a) of the Rome Statute.Footnote 54 Moreover, it is clear that plain meaning only exists ‘in context’, and that the provisions of a treaty cannot be artificially dissected out of the whole and examined. The ICC Appeals Chamber has recognized this, noting that its task is to examine the ‘sub-section of the law [in question] read as a whole in conjunction with the . . . enactment in its entirety’.Footnote 55 In addition, ‘the interpreter must fairly decide which articles are coordinate to the interpretive exercise, and which ones are not’ so as to avoid creating ambiguity via structural readings ‘in which the subject clause is made nonsensical by parallel readings with irrelevant provisions’.Footnote 56
Finally, as recognized in the Vienna Convention, the Court should only move beyond the text when a plain reading of a provision leaves an ‘ambiguous or obscure’ meaning or leads to a result that is ‘manifestly absurd or unreasonable’.Footnote 57 Where this is considered to be the case, judges ought to explicitly detail their reasoning.
2.3.2. Canon 2: Provisions should be construed to be faithful to the object and purpose of the ICC Statute, consistent with the legality principle embodied in Article 22(2)
Article 31 of the Vienna Convention requires not only that a treaty provision be given its ‘ordinary meaning’ but that this meaning should read according to the principle of ‘good faith’, and examined in light of the treaty's ‘object and purpose’.Footnote 58 While seemingly clear, this principle of treaty interpretation is considerably more ambiguous in application, for different constituencies – or different states – may harbour conflicting views regarding the ‘object and purpose’ of a given treaty. With the Rome Statute, in particular, there is a certain tension between the purpose of the Court, ‘to put an end to impunity for perpetrators of [atrocity] crimes and thus to contribute to the prevention of such crimes’ (the object and purpose of the Court set forth in the preamble),Footnote 59 and the legality principle in Article 22(2). As noted in Section 2.2 above, many provisions of the Rome Statute are susceptible to liberal overreach if they are interpreted according to the ‘object and purpose’ of the Rome Statute. Yet although ‘interpretation cannot camouflage expansion’,Footnote 60 neither should judges craft extraordinarily rigid understandings of the Statute's more open-ended or debatable provisions, but rather interpret these in a manner consistent with the object and purpose of the Statute.
2.3.3. Canon 3: Where the meaning of a particular provision remains ambiguous or obscure or leads to a result that is manifestly absurd or unreasonable, the travaux préparatoires may be consulted
Recourse to the travaux préparatoires may be useful in resolving ambiguities in the text, but ‘inherent weaknesses’ in this source must be realized.Footnote 61 As Bederman notes, ‘treaty-making bears little relation to the legislative process’, whereby votes taken at diplomatic conferences can be ‘misleading as to the consensus that actually exists around certain provisions’.Footnote 62 The International Court of Justice, and the Permanent Court before it, has generally ‘refused to resort to preparatory work if the text is sufficiently clear in itself’.Footnote 63 At the same time, the Court has referred to a treaty's preparatory work ‘to confirm a conclusion reached by other means’.Footnote 64 This suggests that ICC judges may examine the negotiating history of the Rome Statute, but with an eye towards confirming understandings of the text arrived at through less problematic means of analysis. The danger of relying upon individual statements (which may represent only the views of one or two governments) necessitates that the Court ‘must look for a more subtle expression of consensus’.Footnote 65 This can be achieved by focusing less on statements made by individual delegates, and ‘more on consensus explications of documents and the actual evolutionary process of drafting a particular treaty clause’.Footnote 66 As the International Court of Justice famously noted in the Lotus case, it is relatively easy to find arguments on both sides of an interpretative question in the travaux préparatoires of a given treaty provision, leading it to declare that ‘there is no occasion to have regard to preparatory work if the text of a convention is sufficiently clear’.Footnote 67 To this end, emphasis should be placed upon using the travaux préparatoires to confirm or reject a textual meaning, rather than finding a new one.
2.3.4. Canon 4: If gaps remain in the interpretation of a particular provision, the Court should look to Article 21(1)(b) sources of law, including the jurisprudence of the ad hoc international criminal tribunals, and, failing that, Article 21(1)(c) sources
As described in Section 2.1 above, Article 21(1) sets forth a hierarchical listing of the sources of law that can be used by the Court. Treaties, customary international law and general principles of law are, in theory, secondary to the textual analysis described in the first three canons listed above. At the same time, the analysis of a particular provision will necessarily proceed more holistically than the linear progression set forth in Article 21(1). Nonetheless, ICC judges should endeavour to explicitly justify why an Article 21(1)(b) or (c) source of law is being relied on. Presumably, the Court should only move beyond the text of the Statute when a purely textual analysis leaves the meaning of a provision ‘ambiguous or obscure’, leads to a result that is ‘manifestly absurd or unreasonable,’Footnote 68 or where the Statute is simply silent on an important question (the meaning of ‘participate actively in hostilities’ in Articles 8(2)(b)(xxvi) and 8(2)(e)(vii) of the Elements of Crimes as interpreted in Lubanga comes to mind).Footnote 69 Where resort to customary international law or general principles of law is deemed necessary, the Court should itself genuinely engage in this comparative legal analysis and avoid relying heavily on secondary sources and commentary or law emanating from one particular country.Footnote 70 The need to fill gaps is likely to be a frequent occurrence given the paucity of definitions in the ICC Statute itself and the constant stream of unforeseen questions the Court has faced since its inception, and it will be important for the Court to develop a consistent and credible methodology to do so. The jurisprudence of the ad hoc international criminal tribunals was explicitly based upon customary international law, and may therefore be a reserve of useful interpretative doctrine for ICC judges. Of course, to the extent that the jurisprudence of the tribunals was explicitly directed at interpreting their own statutes, for example, the elaboration of the doctrine of JCE, and where the ICC arguably has an explicit provision on the same subject matter (Article 25(3)(d)), ICC judges should look first to the text of the Rome Statute (Canons 1 and 2) rather than unquestioningly incorporating tribunal jurisprudence. However, in interpreting Article 25(3)(d), it would surely be relevant to consider the elaboration of JCE at the ad hoc tribunals in considering the meaning of that provision, since the concepts are clearly related.Footnote 71
2.3.5. Canon 5: All provisions should be construed with the objective of protecting the rights of the accused and ensuring that the application of the Statute is consistent with internationally recognised human rights
What makes interpreting the Rome Statute fundamentally different from many other treaties is that the rights of an accused are at stake. International criminal law is still a relatively new field of law and gap-filling is inevitable. At the same time, making unjustified interpretative leaps in order to ‘end impunity’ risks destroying the legitimacy of the endeavour by subjecting individuals to definitions of crimes that could not possibly have been foreseen. In this context, it is of utmost importance that provisions of the Statute affecting the accused's rights and liability be strictly construed to give a meaning that is predictable and reflective of the basic tenets of domestic criminal law systems. Most importantly, any interpretation must be made with an awareness of the fact that an accused cannot be found liable for conduct which was not a crime at the time of its commission,Footnote 72 meaning that retroactive criminalization via creative interpretations of the Statute must be avoided.
Article 21(3) underscores the fundamental importance the drafters of the Rome Statute attached to maintaining human rights protections in the application of the Statute, not only as regards the accused but also as regards victims and witnesses.Footnote 73 Thoughtful adherence to the admonition of this provision will allow the ICC to serve as a model for the fair administration of justice by states, and enhance not only the combat against impunity, but the building of a human rights culture in the national jurisdictions that will hear most of the cases tried in the aftermath of mass atrocities.
2.3.6. Canon 6: The interpretation adopted should enhance judicial efficiency and the effectiveness of the ICC trial system, without compromising the values expressed in Canon 5
Overly complex or creative interpretations of the Rome Statute are bound to result in disagreement and appeals. International criminal law is, by virtue of its infancy, an unsettled field of law. Judges should seek to interpret the Rome Statute in a manner that does not promote litigation that is unlikely to be fruitful, and should work to promote the efficiency and transparency of the ICC trial process without sacrificing the rights of the accused. In particular, the accused has a right to ‘be tried without undue delay’.Footnote 74 Thus, like the ICTY, ICC judges should consider the need to permit charging in the alternative, not just of crimes but of differing modes of liability under Articles 25 and 28.Footnote 75 This is perhaps one of the most difficult and sophisticated tasks they have been asked to accomplish as the Rome Statute includes procedural innovations not known either to national criminal justice systems or the ad hoc international criminal tribunals (like the confirmation of charges hearing required by Article 61). This might be the area in which judges are freest to fill gaps and interpret the Statute in a purposive manner (subject of course to Article 22(2)). Indeed, without taking sides in the debate between the majority and the dissent in the Katanga Footnote 76 case regarding the appropriateness of recharacterizing Katanga's mode of individual criminal responsibility in that case, in many national jurisdictions and in the ad hoc international criminal tribunals, it is perfectly appropriate to plead modes of liability in the alternative, and allow the court to choose the proper characterization of the accused's criminal participation after hearing the evidence at trial.Footnote 77 Had this been permitted in Katanga, it would have likely avoided both the delay occasioned by the recharacterization of his mode of liability and the subsequent controversy over whether he had been granted a fair and impartial trial.
2.3.7. Canon 7: The interpretation of a particular provision should enhance the expressive and normative function of international criminal law by rendering it transparent and comprehensible and reducing opportunities for fragmentation
Finally, we propose that simplicity, clarity, transparency, and fairness should be the lodestar for ICC Statute interpretation. As will be shown, the breathtaking complexity of the ICC's jurisprudence on modes of liability makes it difficult for even the specialist to understand, and has given rise to an extraordinary number of academic articles debating the meaning (and appropriateness) of the case law. Even more problematic, as it now stands, the ICC is issuing arrest warrants against heads of state like Muammar Gaddafi (Libya) and Laurent Gbagbo (Côte d’Ivoire) as ‘indirect co-perpetrators’ of crimes committed by their own forces, suggesting they are somehow less responsible than they would be if they were ‘direct’ perpetrators.Footnote 78 As one commentator pithily observed, ‘when the ICC indicts former President Laurent Gbagbo as an indirect co-perpetrator, the BBC places the mode of liability in parentheses to mark the technocratic reference to legalese that only a small élite (who has no personal stake in the injustice) is likely to understand’.Footnote 79 This is a real problem for the Court. It is unlikely to promote the kind of normative development needed to promote international criminal justice, and does not truly express the outrage felt that a head of state has presided over the commission of atrocities against civilians.
The ICC should see itself as the permanent embodiment of the customary international law principles established in and by the ad hoc international tribunals except insofar as the Statute clearly directs it in a different direction. This will ensure the universality of the ICC Statute, which applies (through the possibility of Security Council action) to individuals living in Rome State party states and non-state party states, enhance its legitimacy, and help to strengthen and consolidate rather than fragment the application of international criminal law. The Court has occasionally suggested that it cannot look at customary international law established in the ad hoc tribunals;Footnote 80 however, unless its own Statute so requires, maintaining the coherency of international criminal law is vitally important and should be a goal embraced by the judges at the ICC, just as it has been embraced by other international courts and tribunals, such as the International Court of Justice in Bosnia v. Serbia, for example, which extensively cited the jurisprudence of the ICTY in determining whether genocide had been committed in Bosnia. Footnote 81
It is worth noting that one can find very few examples of international courts and tribunals purposefully deciding to fragment international criminal law by opting for differing, rather than uniform, interpretation of their texts.Footnote 82 Jonathan Charney observed this phenomenon in his classic Hague Academy Lecture, noting that many international courts and tribunals share a ‘coherent understanding’ of the international law they have been charged with applying.Footnote 83
3. Evolution of the concept of individual criminal responsibility and the Article 25 debate
Before considering the application of the seven canons set forth above to Article 25(3) (Section 4), it is useful to consider the evolution of this provision under customary international law, in the ad hoc international criminal tribunals (Section 3.1), and during the negotiation of the Rome Statute (Section 3.2). Finally, we turn to the difficulties engendered by Article 25(3)'s complexity in the early jurisprudence of the ICC (Section 3.3).
3.1. Customary international law and the ad hoc tribunals
The principle of individual criminal responsibility means that ‘no one may be held accountable for an act he has not performed or in the commission of which he has not in some way participated, or for an omission that cannot be attributed to him’.Footnote 84 Applying this seemingly simple proposition to those accused of international crimes has challenged international criminal courts and tribunals since the end of the Second World War.
References to the principle of individual criminal responsibility appeared in some international legal documents prior to the post-Second World War tribunals.Footnote 85 However, it was the Charters governing the International Military Tribunal (IMT) at Nuremberg and the International Military Tribunal for the Far East (IMTFE) at Tokyo that first set out modes of criminal responsibility, albeit ‘scattered throughout the text’.Footnote 86 There is a general consensus that the IMT and IMTFE Charters and case law adopted a unitary model of attribution, which did not strictly distinguish between the perpetration of a crime (principal liability) and participation in a crime committed by a third person (accessorial liability), and did not suggest that the degree of an accused's culpability depended upon the form of criminal responsibility under which he was charged.Footnote 87 This is particularly evident in Article II(2) of Control Council Law No. 10, which provides:
Any person . . . is deemed to have committed a crime as defined in paragraph 1 of this Article, if he was (a) a principal or (b) was an accessory to the commission of any such crime or ordered or abetted the same or (c) took a consenting part therein or (d) was connected with plans or enterprises involving its commission or (e) was a member of any organization or group connected with the commission of any such crime or (f) with reference to paragraph 1 (a) if he held a high political, civil or military (including General Staff) position in Germany or in one of its Allies, co-belligerents or satellites or held high position in the financial, industrial or economic life of any such country.Footnote 88
As a textual matter, this provision suggests that all forms of criminal participation were forms of ‘commission’ before US military tribunals, and that any distinction between principal and accessory was ‘irrelevant’ in practice.Footnote 89
In the 1990s, the Statutes of the ICTYFootnote 90 and ICTRFootnote 91 were adopted. Article 7(1) of the ICTY Statute and Article 6(1) of the ICTR Statute state that ‘[a] person who planned, instigated, ordered, committed or otherwise aided and abetted in the planning, preparation or execution of a crime referred to in [the Statute], shall be individually responsible for the crime’. The initial understanding of these provisions suggested they follow a unitary perpetrator model because they put ‘the commission of the crime on the same level and within the same category as planning, instigating, ordering, or otherwise aiding’.Footnote 92 The subsequent elaboration of the draft code of offences against the peace and security of mankind by the International Law Commission (ILC) included the modes of liability found in earlier instruments as well as the Statutes of the ICTY and ICTR separated out into sub-paragraphs, but was silent on the question of ‘unitary’ versus ‘differentiated’ modes of liability, suggesting either no view as to the same, or that customary international law continued to regard all those who commit crimes under international law as ‘perpetrators’ regardless of the form of their criminal participation. Footnote 93
Recent case law from the ICTY and ICTR suggests a departure from the ‘unitary’ perspective previously extant under customary international law, attributing greater culpability to ‘principals’ than ‘aiders and abettors’. This is evidenced by cases such as Krstić, in which the ICTY Appeals Chamber lowered the accused's sentence based upon its conclusion that he was an ‘aider and abettor’ rather than guilty of ‘complicity’ as a principal.Footnote 94 Although this issue is not entirely settled at the ICTY,Footnote 95 and discussion of the issue at the ICTR has been notably more limited, given the express reliance on customary international law by the ad hoc tribunals, it may be relevant either as evidence of an evolution in customary international law (or may simply be an interpretation of the ICTY/ICTR Statutes).
The ICTY and the ICTR treat ‘common plan’ liability as a primary form of commission, having developed the doctrine of joint criminal enterprise (JCE) to do so. This form of criminal responsibility was introduced by the ICTY Appeals Chamber in the Tadić case as a ‘gap filling’ measure required by the failure of the ICTY/ICTR Statute drafters to include common plan or conspiracy in the modes of liability available to those tribunals, bases of liability that were clearly present in earlier international criminal law instruments.Footnote 96 The ICTY Appeals Chamber has held that the JCE doctrine constitutes customary international law.Footnote 97 Although most commentators agree, there is no doubt that the elaboration of this form of liability by the tribunals’ judiciary was one of its most controversial efforts, particularly as regards the most extended form of JCE liability, which has been widely criticized in the literature as unsupported in customary international law and insufficiently attentive to the principles of nullum crimen sine lege and individual culpability.Footnote 98 In other words, although many would agree that the gap-filling nature of this doctrine places it's elaboration in the ‘judicial creativity’ category,Footnote 99 failure to carefully adhere to a clear methodological approach in doing so has some wondering whether this wasn't more like judicial activism.Footnote 100
Scholars are divided as to the import of this evolution at the ad hoc tribunals. Although ‘the practical relevance of distinguishing between principals and accessories is limited’, it is arguable that the distinction in international criminal law has ‘gained importance in recent years’ at least insofar as sentencing is concerned.Footnote 101 In the views of some, classifying particular forms of criminal participation as fitting into one mode of participation or another, ‘serves a descriptive and conceptual or classificatory purpose only . . . devoid of any relevance as far as sentencing is concerned’Footnote 102 and more recently, the Special Court for Sierra Leone rejected the principal/accessory distinction for the purposes of sentencing in the Charles Taylor case, sentencing him to 50 years for ‘aiding and abetting’. In finding that aiding and abetting did not generally warrant a lesser sentence, the Appeals Chamber held that the Court's Statute does not refer to or clearly establish a hierarchy of criminal participation. To find that such a hierarchy exists would, in the Appeal Chamber's view, be contrary ‘to the essential requirement of individualisation that derives from the mandate of the Court, principles of individual criminal liability and the rights of the accused’.Footnote 103 So while some may argue that that the differentiated approach to attributing liability is taking hold in international criminal law, it appears that the law is still in a state of flux.
3.2. Article 25 of the Rome Statute: a Rorschach blot?
Article 25 sets out a more comprehensive and detailed framework of liability than predecessor instruments and was the result of lengthy negotiations. Sub-paragraph (3)(a) introduces the concept of perpetration by commission of a crime by a person as an individual, jointly with another or through another person, while sub-paragraphs (3)(b)–(c) set out a variety of other forms of liability including ordering, soliciting, inducing, aiding and abetting. Sub-paragraphs (d), (e), and (f) provide for contributing to the commission or attempted commission of a crime by a group, incitement to genocide, and attempt. Per Saland, who chaired the Working Group on General Principles of Criminal Law throughout the ICC negotiations, has noted that the provision
. . . posed great difficulties to negotiate in a number of ways. One problem was that experts from different legal systems took strongly held positions, based on their national laws, as to the exact content of the various concepts involved. They seemed to find it hard to understand that another legal system might approach the issue in another way: e.g., have a different concept, or give the same name to a concept but with a slightly different content.Footnote 104
While Article 25 has its origins in the work of the ILC on the Code of Crimes Against the Peace and Security of Mankind, later drafts more closely representing the final version of Article 25 were proposed by ‘an informal group representing various legal systems’.Footnote 105 Indeed, the final version draws on various sources of national criminal law, including, but not limited to, German law and international treaty provisions.Footnote 106 The vast and divergent literature on Article 25 suggests it is a sort of ‘Rorschach blot’, in which scholars tend to see and read into the provision their own experience and understanding of criminal liability, based on their national legal system (including, admittedly, these authors).Footnote 107
In our view, the travaux préparatoires do not demonstrate a clear preference for a strict principal/accessory distinction or suggest that the modes of liability listed in Article 25 were intended to be ranked hierarchically. Indeed, although some scholars have argued that Article 25 embodies both a differentiated and hierarchical approach to criminal participation,Footnote 108 others disagree, noting that ‘it is difficult to find an unambiguous answer’ to the question of what model of perpetration and participation it adopts.Footnote 109 Article 25(3) is likely a consensus provision that lacks a strong and logically cohesive theoretical underpinning of the kind that can be found in domestic jurisdictions.Footnote 110 In particular, the Statute drew the language on common plan liability in paragraph 3(d) from the International Convention for the Suppression of Terrorist BombingsFootnote 111 to avoid divisive discussions about incorporating the concept of conspiracy, which was rejected by most states parties.Footnote 112
3.3. Emerging divisions at the ICC
The ICC has struggled to give substantive content to Article 25 from its earliest days, with sub-paragraph (a) (‘commission’) receiving the most attention. As will be shown, early pre-trial chamber decisions set the course for the ICC's jurisprudence on Article 25, often with very little reasoning. There is, however, growing discontent within the Court regarding both the substantive law on Article 25 and the interpretative methodology used to justify it.
3.3.1. Pre-trial chamber decisions
In an early arrest warrant decision in the Lubanga case in 2006,Footnote 113 Pre-Trial Chamber I stated – without citing any authorities in support — that there is a distinction between (i) the commission stricto sensu of a crime by a person as an individual, jointly with another or through another person within the meaning of Article 25(3)(a) of the Statute, and (ii) the responsibility of superiors under Article 28 of the Statute and ‘any other forms of accessory, as opposed to principal, liability provided for in article 25(3) (b) to (d) of the Statute’.Footnote 114 Curiously, it did so as a means of interpreting the word ‘committed’ in Article 58(1), distinguishing ‘committed’ in the Article 58(1) sense from ‘committed stricto sensu’ in Article 25(3)(a).Footnote 115 It adopted the view that ‘the concept of indirect perpetration [. . .], along with that of co-perpetration based on joint control of the crime referred to in the Prosecution's Application’, was ‘provided for in Article 25(3)(a) of the Statute’.Footnote 116
Subsequently, Pre-Trial Chamber I affirmed the principal/accessory distinctionFootnote 117 as well as the incorporation of the ‘control of the crime’ theory, for which it cited the work of three scholars and the separate opinion of Judge Schomburg in the Gacumbitsi case, decided by the ICTR.Footnote 118 Rejecting the argument that joint criminal enterprise liability was incorporated into Article 25(3)(a), the judges opted instead for the ‘control of the crime’ approach, holding
that principals to a crime are not limited to those who physically carry out the objective elements of the offence, but also include those who, in spite of being removed from the scene of the crime, control or mastermind its commission because they decide whether and how the offence will be committed.Footnote 119
This interpretation was challenged by defence lawyers as beyond the ‘clear terms’ of the Statute and ‘not supported by customary international law, or general principles of law derived from legal systems of the world’.Footnote 120 No doubt, defence counsel reading Judge Schomburg's opinion in Gacumbitsi were worried about the potential expansion of attribution to perpetrators far from the scene of the crime given his statement that this idea ‘suits the needs . . . of international criminal law particularly well . . . [as] a means to bridge any potential physical distance from the crime scene of persons who must be regarded as main perpetrators’.Footnote 121 It was supported, however, both by the legal representatives of the victims and the prosecutor who argued that co-perpetration under Article 25(3)(a) best represented the criminal responsibility of the accused.Footnote 122
Unlike JCE, which had antecedents in the Nuremberg precedent, the control theory is new to international criminal proceedings. Premised on German legal doctrine and the writings of Claus Roxin, a renowned German legal scholar,Footnote 123 it represents neither a purely subjective nor objective approach for distinguishing between principals and accessories.Footnote 124 Under this third conception, co-perpetration ‘is rooted in the principle of the division of essential tasks for the purpose of committing a crime between two or more persons acting in a concerted manner’.Footnote 125 All participants ‘share control because each of them could frustrate the commission of the crime by not carrying out his or her task’.Footnote 126 The Pre-Trial Chamber in Lubanga asserted that the control theory is applied in many legal systems,Footnote 127 but that does not appear to be the case,Footnote 128 and in any event, it neither conducted a survey of these systems nor attempted to establish that the theory constituted a ‘general principle of law’ under Article 21(1)(c) of the Rome Statute. Finally, the Pre-Trial Chamber did not explicitly reason that the adoption of this theory was necessary to fill a lacuna in the text of the Statute.
Following this decision, Pre-Trial Chamber I in the Katanga/Ngudjolo Confirmation DecisionFootnote 129 examined the notion of committing a crime ‘through another person’ (‘indirect perpetration’) under Article 25(3)(a). The Pre-Trial Chamber confirmed the adoption of the control theory in Katanga, citing a range of secondary material to support the argument that it applied in a number of legal systems.Footnote 130 Noting that ‘article 25(3)(a) uses the disjunctive “or”’, the Pre-Trial Chamber reasoned that
. . . through a combination of individual responsibility for committing crimes through other persons together with the mutual attribution among the co-perpetrators at the senior level, a mode of liability arises which allows the Court to assess the blameworthiness of “senior leaders” adequately.Footnote 131
It concluded:
An individual who has no control over the person through whom the crime would be committed cannot be said to commit the crime by means of that other person. However, if he acts jointly with another individual — one who controls the person used as an instrument — these crimes can be attributed to him on the basis of mutual attribution.Footnote 132
The result was the controversial concept of ‘indirect co-perpetration’, ‘a type of “perpetrator behind the perpetrator” liability based on control over a hierarchical organization (Organisationsherrshaft).’Footnote 133
The reasoning and conclusions drawn from the Lubanga and Katanga/Ngudjolo confirmation decisions have been repeated in successive arrest warrant and confirmation of charges decisions with relatively little new analysis or critique, particularly in relation to interpretative methodology.Footnote 134
3.3.2. The Lubanga trial judgment
The majority in Lubanga devoted nearly thirty-five pages of its judgment to the question of the mode of liability charged. It predicated its analysis of Article 25(3)(a) on a policy argument: that the modes of liability set out in the Statute must be interpreted in a way that allows for ‘properly expressing and addressing’ responsibility for crimes under the jurisdiction of the Court.Footnote 135 It suggested that the words ‘jointly with another’ in Article 25(3)(a) require that (1) at least two individuals are involved in the commission of the crime and (2) an agreement or common plan exists between them.Footnote 136 The majority concluded that the common plan must include ‘a critical element of criminality, namely that, its implementation embodied a significant risk that, if events follow the ordinary course, a crime will be committed’,Footnote 137 and that the accused must provide an ‘essential contribution to the common plan’.Footnote 138 However, he need not ‘be present at the scene of the crime, as long as he exercised, jointly with others, control over the crime’.Footnote 139 The majority differentiated between Article 25(3)(a) and Articles 25(3)(b)–(d), suggesting that the former embodies a form of principal liability, while paragraphs (b) to (d) attribute accessorial liability, arguing that its interpretation ‘allows for the different degrees of responsibility to be properly expressed and addressed’.Footnote 140
Judge Fulford sharply disagreed. He argued that ‘the test laid down by the Pre-Trial Chamber is unsupported by the text of the Statute and it imposes an unnecessary and unfair burden on the prosecution’.Footnote 141 Indeed, in Fulford's view, paragraphs (a) to (d) of Article 25(3) were not meant to create a hierarchy of mutually exclusive forms of liability, as the majority contended. Rather, they were intended to anticipate the numerous and interconnected ways that a person might incur criminal responsibility.Footnote 142 Moreover, he rejected the notion that Article 25(3) creates a ‘hierarchy of seriousness as regards the various forms of participation in a crime’.Footnote 143 He noted that such an approach might well be warranted within the German legal system (upon which the majority based its analysis) in which the sentencing range is determined by the mode of liability, but is inappropriate at the ICC where sentencing is not tied to any distinction between principals and accessories.Footnote 144 Finally, he suggested that the expression ‘commits . . . jointly’ self-evidently necessitates a ‘sufficient meeting of the minds’ and that there is no evidence that the Statute requires proof that the crime would not have been committed but for the accused's contribution,Footnote 145 thereby avoiding ‘a hypothetical investigation as to how events might have unfolded without the accused's involvement (which is necessary under the ‘essential contribution’ formulation)’.Footnote 146
3.3.3. The Ngudjolo trial judgment
On 18 December 2012, Trial Chamber II acquitted Mathieu Ngudjolo Chui, the alleged former leader of a Congolese rebel group involved in the Ituri conflict, of war crimes and crimes against humanity.Footnote 147 The majority did not address the proper interpretation of Article 25(3)(a) in its judgment. However, Judge Van den Wyngaert did in her separate opinion,Footnote 148 which outlined why, in her view, the control theory should be rejected by the Court. She advanced several reasons for her conclusion, which may be summarized as follows. First, she argued that the control theory is not consistent with Article 22(2) of the Statute (requiring that a definition of a crime be strictly construed) and the ordinary meaning of Article 25(3)(a). Second, she argued that the premise justifying incorporation of the control theory, that there is an alleged hierarchy in the modes of liability, is incorrect. Third, she contended that the control theory's treatment of the common plan as an ‘objective’ as opposed to ‘subjective’ element unduly focuses on the accused's link to the common plan as opposed to the crime; and finally, she found there to be no legal basis for the ‘essential contribution’ requirement imposed by the majority.Footnote 149 She concluded that ‘perpetration through another person can [not] be equated to control over an organisation and . . . that the notion of “indirect co-perpetration” has no legal basis in the Statute’.Footnote 150
Judge Van den Wyngaert, like Judge Fulford, adopted a ‘plain meaning’ approach to Article 25(3), with some distinctions. She devoted considerable attention to justifying the interpretative methodology that guided her conclusion:Footnote 151 that for the purposes of joint perpetration under Article 25(3)(a), the contribution must be direct, not essential (as the preceding cases suggest), because ‘[o]nly those individuals whose acts made a direct contribution to bringing about the material elements can thus be said to have jointly perpetrated the crime’.Footnote 152 A ‘direct contribution’, according to Van den Wyngaert, ‘is an immediate impact on the way in which the material elements of the crimes are realised’.Footnote 153 It does not, however, ‘necessarily require the physical presence of the joint perpetrator on the scene of the crime and may, depending on the circumstances of the case and the nature of the crime charged, include certain forms of planning and coordination’.Footnote 154
Judge Van den Wyngaert also rejected the Trial Chamber's interpretation of ‘commission . . . through another person’, again opting to give the words their plain meaning. She argued that the concept of ‘indirect co-perpetration’, reached by combining joint perpetration and perpetration through another person, is a radical expansion of Article 25(3)(a) and inconsistent with Article 22(2) of the Statute (incorporating the principle of legality).Footnote 155
3.3.4. The Katanga trial judgment
On 7 March 2014, the majority of Trial Chamber II found Germain Katanga guilty of complicity within the meaning of Article 25(3)(d), of one crime against humanity and four war crimes committed on 24 February 2003 during an attack on the village of Bogoro, in the Ituri district of the Democratic Republic of the Congo (DRC).Footnote 156 The majority confirmed the ‘control of the crime theory’ as consistent with the application of Article 25(3)(a), quoting extensively from Roxin regarding its application. At the same time, it suggested that it was not the only possible understanding of that provision, noting that ‘elle estime qu’il n’y a pas lieu de faire de cette théorie un élément constitutif incontournable de la commission par l’intermédiaire’. Footnote 157 It also found that that two forms of responsibility, ‘auteur’ and ‘complice’ (presumably translated as ‘principal’ and ‘accessory’) is inherent in Article 25(3), but, unlike all previous cases, rejected that the modes of liability in Article 25(3)(a) are ordered hierarchically.Footnote 158 Finally, it simplified the test for indirect co-perpetration to a certain degree, but in so doing added yet another layer of potential confusion to the understanding of Article 25(3).Footnote 159 In sum, the majority judgment represented a significant divergence from prior ICC case law and heightens the case for revisiting Article 25 in the Lubanga Appeal. Judge Van den Wyngaert, in her Minority Opinion, also distanced herself from the majority's approach to recharacterizing the charges in the caseFootnote 160 and reiterated her position outlined in Ngudjolo regarding the correct interpretation of Article 25.Footnote 161
4. Application of the seven canons to Article 25(3)
Our analysis suggests that the majority opinions in Lubanga and Katanga/Ngudjolo (on confirmation) are methodologically flawed both as regards the reading of a hierarchical differentiated model of criminal responsibility into the text of Article 25(3) (Section 4.1) and as regards the importation of the control of the crime theory into Article 25(3)(a) (Section 4.2). Both holdings are inconsistent with Canon 1 (ordinary meaning), Canon 2 (object and purpose as constrained by the principle of legality), and Canon 3 (consultation of travaux préparatoires). They are not justified by a plain reading of the text, do not appear driven by the object and purpose of the Statute, may be inconsistent with the legality principle, and are not supported by a reading of the ICC's travaux préparatoires, which are at best inconclusive (suggesting they should not guide the Court's opinions in any event). Second, and perhaps even more problematically, these decisions and the preceding opinions upon which they are based made no effort to comply with the hierarchy of sources set forth in Article 21 (text followed by treaties, customary international law, and general principles of law) (Canon 4) and, to the extent they do consider the question, do not analyse those sources in a manner appropriate to their consideration before an international tribunal. It is possible that the superposition of a hierarchy upon the text of Article 25(3) complies with Canon 5, although it is equally plausible that the complexity introduced by the majority might work to the disadvantage of the accused and impede rather than promote the protection of human rights. This is a much greater risk as regards importation of the control theory into Article 25(3). Finally, both the establishment of a hierarchy of offences and the importation of the control of the crime theory into Article 25(3) has arguably led to a decrease in efficiency, the imposition of additional burdens on the prosecutor, fragmented international criminal law by imposing a sharp break between the practice of the ad hoc international criminal tribunals and the ICC, and created an extraordinarily complex and esoteric understanding of Article 25(3) that renders international criminal law obscure and difficult to understand (violating Canons 6 and 7). An explanation of these conclusions follows.
4.1. A hierarchy of modes of liability
Reading a hierarchy into the text of Article 25(3) seems clearly to breach six out of seven of our canons, and perhaps all seven. A plain reading of the text of Article 25, in accordance with Canons 1 and 2, does not suggest the modes of liability are intended to be hierarchal in order of gravity.Footnote 162 Certainly, there is no evidence that ‘ordering’ (in 25(3)(b)) is less culpable than ‘committing’ in 25(3)(a) or that ‘common plan liability’ in 25(3)(d) is less culpable than ‘committing’ or ‘ordering’. It is possible to read 25(3)(c) (aiding and abetting) as suggesting some differentiation of culpability, but it is equally plausible to read this provision as not mandating any differentiation. It is perhaps worth observing that where the negotiators intended a provision to embody a hierarchy, they did so quite explicitly, as shown by Article 21 itself which employs the language, ‘in the first place’, followed by ‘in the second place’, followed by ‘failing that’.Footnote 163 According to the contemporaneous histories of Article 25 that have been published, it was drafted to reflect a variety of overlapping modes of liability, incorporating diverse national and international models. An examination of the travaux préparatoires, while not particularly instructive, confirms this view and indicates that the states parties intended to create a consensus provision that reflected most common forms of criminal liability throughout the world, rather than create new forms of liability specifically suited to atrocity crimes.Footnote 164
Likewise, the adoption by the ICC of the principle/accessory distinction is similarly not well supported, although we do not go so far as to say it is forbidden by the text of Article 25.Footnote 165 It translates poorly into the French text of the Statute (which is unsurprising since the concept is not really present in the same way in French law) and had the negotiators wished to include the words ‘principal’ or ‘accessory’ in the text, they could have done so.Footnote 166 Finally, even had they included the terminology of ‘principal’ and ‘accessory’ in the ICC Statute, it is worth observing that the Nuremberg judgment and the text of Control Council Law 10 embodied a unitary approach to culpability while still employing the terms ‘principal’ and ‘accessory’. While this terminology appears in the jurisprudence of the ad hoc international criminal tribunals, that practice is far from uniform, and could not, in any event be adopted at the ICC unless it forms part of customary international law or general principles, an analysis that the ICC has not undertaken except to say that it is ‘implicit’ in the text of Article 25(3)Footnote 167 (in violation of Canon 4).
It is also possible that attributing greater gravity to particular modes of liability over others risks breaching the rights of the accused (contrary to Canon 5). As the Charles Taylor case noted,Footnote 168 the seriousness of ones’ offending must be judged on the facts. Sentencing on the basis of the mode of liability charged risks either attributing a sentence too high or too low depending on the circumstances. Indeed, Article 76 of the Statute, read in conjunction with Rule 145, make clear that in determining sentences, the Court shall consider all mitigating and aggravating factors and ‘consider the circumstances both of the convicted person and of the crime’, including the ‘nature of the unlawful behavior and the means employed to execute the crime’ and ‘the degree of participation of the convicted person’.Footnote 169 Thus the ICC statutory scheme clearly places assessment of culpability at the end of the trial proceedings, once the evidence has been adduced and tested at trial, and not at the beginning in determining the accused's mode of liability. It is certainly more efficient to assess culpability after the evidence has been heard, than make guesses about it ex ante. Attempting to establish with precision the ‘one’ mode of liability attributable to an accused's conduct at the confirmation stage of the proceedings may be inefficient and even inappropriate given that limited evidence is available at that stage of the proceedings, contrary to Canon 6. It would seem preferable to permit the case to move forward alleging the forms of participation charged (committing, ordering, providing assistance, etc.), and then select the appropriate mode of participation established by the evidence presented at trial. Finally, reading a hierarchy or principal/accessory distinction into Article 25 does not improve the expressive or normative function of international criminal law (Canon 7).Footnote 170 Lay readers are unlikely to recognize differences between modes of liability charged and are more likely to view the sentence given and description of the requisite conduct as reflective of the seriousness of the crime.Footnote 171 Although some authors have suggested that a hierarchical understanding of Article 25 would promote ‘fair labeling’ and thereby have expressive value, the language and theory the Court has adopted is stunningly complex, causing scholars the world over to argue over its meaning and appropriateness, and suggesting that the educated lay reader is unlikely to understand much at all. The ICC should be seeking simplicity and universality in the labels it attaches to offenders and the prevailing interpretation of Article 25(3) does not achieve this.
4.2. The importation of the control theory
The importation of the control theory is perhaps even more problematic than the insinuation of a hierarchy into the text of Article 25(3). As Judges Fulford and Van den Wyngaert demonstrated, a plain reading interpretation of Article 25, in accordance with Canons 1 and 2, is indeed possible, albeit contestable. Yet the ICC's pre-trial chambers simply never identified why a plain reading of the text was unworkable. The prevailing interpretation of Article 25 at the ICC also exemplifies how an ‘object and purpose’ approach has been employed to justify an otherwise improper incorporation of an isolated legal theory in violation of the legality principle (Canon 2). As Judge Van den Wyngaert noted, ‘[e]ven if the “fight against impunity” is one of the over-arching raisons d’être of the Court which may be relevant for the interpretation of certain procedural rules, this cannot be the basis for a teleological interpretation of the articles dealing with criminal responsibility’,Footnote 172 suggesting that the inclusion of Article 22(2) in the Statute applies to the definition of criminal responsibility and indeed overrides convention interpretation methods in the VCLT.Footnote 173
Even if, for argument's sake, it is accepted that a gap in the text needed to be filled (Canon 4), the control theory is unlikely to be found either in customary international law or as a ‘general principle’ of law and thus unhelpful as a subsidiary gap-filling source of law.Footnote 174 The travaux préparatoires make no reference indicating the states parties intended that this theory be read into Article 25 (Canon 3).Footnote 175 As previously stated, if the drafting history of the provision indicates anything, it is that Article 25 was drafted to incorporate varying and overlapping modes of liability designed to satisfy the wishes of states parties who came from a raft of different legal systems. The fact that common plan liability, in paragraph 3(d), is drawn from an international treaty is further indicative of an attempt to include a form of common plan liability palatable to most states parties.Footnote 176 The effectiveness and clarity of this particular mode of liability as it is currently drafted can be debated, but it should be for the states parties (not the judiciary) to resolve whether it is appropriately meeting its goals.Footnote 177
Finally, the importation of the control theory may compromise the rights of the accused, hamper judicial efficiency and the effectiveness of ICC proceedings, and does little for the expressive and normative function of international criminal law, contravening Canons 5–7. It is difficult to see how any accused, or many legal scholars for that matter, could have predicted that Article 25 would be interpreted to incorporate an isolated theory of criminal liability, and it is hard to see how it enhances the efficiency and effectiveness of the Court's proceedings to require the prosecutor to choose between an offender's categorization as a ‘direct’ and ‘indirect’ co-perpetrator. Indeed, it is unclear why it is helpful to read the requirement of a ‘common plan’ into the notion of joint commission under Article 25(3)(a) when it is already included as an element in Article 25(3)(d).
Judge Shahabudeen recently observed that both joint criminal enterprise and the control theory are not universally accepted, but argues that the ICTY was being (appropriately) judicially creative when it adopted the doctrine of joint criminal enterprise.Footnote 178 While an argument can and has been made that the ICC is similarly free to adopt the control theory, it remains unclear why the ICC should depart from the significant body of law developed by the ICTY on modes of liability. As we have argued, a plain reading of Article 25 does not justify the importation of either theory, but it is particularly curious as to why the Court has decided to specifically depart from ICTY jurisprudence. In this sense, the prevailing interpretation of Article 25 at the ICC represents a serious fragmentation of international criminal law, and one that may not come with any real benefits.Footnote 179
5. Conclusion
This article suggests that there are sound arguments for rejecting the prevailing interpretation of Article 25(3) at the ICC both as to the incorporation of a differential hierarchy in the sub-paragraphs of Article 25(3)(a) through (d) and its incorporation of the control theory to explain joint commission in Article 25(3)(a). Indeed, it posits that the Court's current approach is fundamentally flawed at a methodological level, and suggests an approach that is consistent with prevailing understandings of treaty hermeneutics. The efforts of the Trial Chamber in Katanga are important and positive steps in the rights direction, but in the views of these authors, does not go quite far enough. This article does not delve into the theoretical specifics of what a plain reading of Article 25 should look like, but the contributions made by Judges Fulford and Van den Wyngaert have provided significant groundwork for this debate,Footnote 180 and we have appended a table that suggests a similar outcome to the one they propose.
The seven canons of ICC interpretation outlined in this paper are designed to provide a basic framework that balances traditional methods of interpretation with the unique characteristics of the Rome Statute. Application of these canons to Article 25 should ideally lead to a simpler, more internationally acceptable and predictable understanding of modes of liability at the ICC, one that does not sharply break with the practice of the ad hoc tribunals, but remains, at the same time, faithful to the provisions of the ICC Statute. This approach departs from the complexity of the Court's current decisions and recognizes that it is not the Court which is tasked with elaborating novel theories (particularly after many accused had already committed their alleged crimes).
The ICC is, unlike the ad hoc tribunals, a permanent institution; the theoretical construction embarked upon in its early cases will be critical to the future effectiveness and legitimacy of international criminal law. The Court must send a clear and unified message to potential perpetrators of atrocities of what constitutes the attribution of criminal liability. The crafting of this message requires simplicity and should attract a broad international consensus. It must also keep in mind that, through the vehicle of the Security Council referral process, the Rome Statute is applied to individuals whose states have not ratified its provisions. The legitimacy of their prosecution depends upon the ICC's Statute remaining consistent with customary international law understandings of international criminality.
When the Rome Statute was negotiated, it was hoped that the blend of the common and civil law traditions incorporated into its provisions would prove a felicitous melange that would protect the rights of the accused, render the trial process efficient and transparent, and protect state sovereignty by filtering out cases that should not be brought.Footnote 181 What the past decade of ICC practice has suggested, however, is that the task of bringing the Statute to life is more arduous than previously expected. In fact, the mix of common and civil law concepts has been difficult to negotiate, and the procedural novelties of the Rome Statute (such as the introduction of the pre-trial phase) have encumbered proceedings considerably. It is hoped that this article, with its suggestion of a methodology that can assist the Court in interpreting ambiguous and open-textured provisions of the Statute's criminal code, will be of assistance as the Court continues its challenging and important work.
Appendix 1: Comparing various approaches to Article 25(3)
