The Identification of Customary Rules in International Criminal Law

This paper aims to examine whether a different methodology has emerged to identify customary rules in the field of international criminal law. For this purpose, this paper briefly touches upon debates regarding customary law as a source and an interpretative aid of international criminal law. It then critically studies the identification methodology of customary law to seek whether a new approach deviating from the classic two-element (State practice and opinio juris) approach is emerging in academia. It also recapitulates some cases of international criminal tribunals to ascertain whether these tribunals have formed a distinct method for custom identification. Finally, it explores the unique characteristics and difficulties in identifying customary rules in international criminal law. It concludes that a different method has not been developed in academia or adopted by tribunals in practice to identify customary rules in international criminal law. The two-element approach still serves as guidance for custom-identification in general, but a flexible application of it is acceptable in specific cases. International practitioners should be cautious in the identification of customary rules in international criminal law, so as to prosecute and punish suspects of international crimes without endangering the principle of legality.


I. Introduction
Customary international law is important for practitioners in international and national courts. Parallel with the movement for the codification of international law, custom and treaties may co-exist on the same subject matter. 1 In this circumstance, a rule that exists in each of these two sources may overlap or conflict with each other or may have identical content, whereas the two sources do not supplant each other and have separate methods of application. 2 In addition, customary law continues to govern the area not stipulated by treaty law. 3 This paper aims to examine the method for ascertainment of customary rules in the field of international criminal law. When we ask the question of how to identify customary law, we refer to a method to ascertain the existence of a customary rule rather than the substance of that rule. In other words, the former deals with the process of identifying whether a customary rule has formed; while the latter concerns the content of a customary rule. 4 The classic approach to identifying custom is to seek sufficient evidence of State practice and opinio juris (two-element approach). The recent work of the International Law Commission (ILC) has also adopted the two-element approach, namely ' a general practice' and ' accepted as law' in the identification of customary international law. 5 It should be noted that, due to an inherent dilemma concerning the A different view, however, is more persuasive. The principle of legality requires that prosecution and punishment be based upon clear provisions of international law at the time the crime was committed. 16 A strict principle of legality contains four derivations: specificity and certainty; non-retroactivity (lex praevia); the ban on analogy (lex stricta); and favouring the suspected or accused person (in dubio pro reo). 17 The rule of specificity and certainty requires the definition of crimes to be sufficiently clear and precise. The rule of non-retroactivity prohibits prosecuting an individual for offences committed before the law that criminalised these conducts came into force as a crime. The rules of the ban on analogy as well as favouring the accused are explicitly enshrined in Article 22(2) of the 1998 Rome Statute of the International Criminal Court (Rome Statute). 18 It stipulates that: '[t]he 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'.
It should first be noted that the difference between treaties and custom in legal certainty is a matter of degree. If the ambiguous attribute of custom would deny its source status, treaties would also be excluded as a source of international criminal law, which is unacceptable. Secondly, international jurisprudence and researchers have upheld that this principle has not been violated if the requirements of foreseeability and accessibility are satisfied. 19 The principle of legality itself serves to limit the interpretation of applicable law, including customary law, instead of excluding custom as a source of international criminal law.
Customary law either as a source of law or as an aid to interpreting written rules is not merely a theory in the field of international criminal law. The Statute of the International Criminal Tribunal for the former Yugoslavia (ICTY) does not stipulate its applicable law. The UN Secretary-General, however, noted that this tribunal should only apply 'rules of international humanitarian law that are beyond any doubt part of customary law'. 20 The drafters' purpose was to limit the ICTY's jurisdiction over crimes existent under customary law so as to avoid violating the principle of legality. 21 In practice, the Appeals Chamber of the ICTY in the Tadić Interlocutory Appeal on Jurisdiction tried to establish a customary rule criminalising serious violations in non-international armed conflicts as war crimes. 22 In another case, the defendant disputed the existence of joint criminal enterprise (JCE) in customary law based on decisions of the Extraordinary Chambers in the Courts of Cambodia. 23 The Trial Chamber of the ICTY rejected this argument and concluded that the Tadić Appeal Judgment established the customary status of JCE. 24   The jurisprudence of other international and internationalised tribunals also supports this viewpoint, 26 although their founding instruments do not directly and explicitly mention customary rules. 27 Customary law has been used to address the challenge to the jurisdiction at the IMT and the Special Court for Sierra Leone (SCSL). 28 In the Nahimana case at the International Criminal Tribunal for Rwanda (ICTR), 29 one defendant appealed against a sentence of 35 years imprisonment by arguing that Article 77 of the Rome Statute provides for 30 years as a maximum possible sentence. 30 The Appeals Chamber rejected this argument. One reason for the Appeals Chamber's decision is that the appellant failed to prove that Article 77 of the Rome Statute was a customary rule in force in 1994. 31 In brief, practitioners could invoke customary rules to argue for or against a rule before these international and internationalised criminal tribunals.
The ICC's applicable law is stipulated in Article 21 of the Rome Statute. 32 Article 21(1)(b) provides that the ICC shall apply '[i]n the second place, where appropriate, applicable treaties and the principles and rules of international law, including the established principles of the international law of armed conflicts'. As Schabas noted, by differentiating the language used in Article 21(1)(b) from the phrase 'general principles of law derived from national laws' in Article 21(1)(c), this expression 'the principles and rules of international law' includes customary law. 33 Other scholars and international authorities also upheld this idea. 34 Under Articles 21(1) of the Statute, customary law is the secondary source of applicable law for the ICC as after the written rules included within the Rome Statute, Elements of Crimes, and the Rules of Procedure and Evidence. 35 For example, as the Pre-Trial Chamber of the ICC in Katanga and Ngudjolo stated: 'since the Rome Statute expressly provides for this specific mode of liability [co-perpetration through another person (indirect co-perpetration)], the question as to whether customary law admits or discards the ''joint commission through another person'' is not relevant for this Court'. 36 The Pre-Trial Chamber in Ruto confirmed this view by arguing that customary law should be of limited applicability within the case because indirect co-perpetration is encompassed by Article 25(3)(a) of the Rome Statute. 37 Additionaly, the reference to ' a crime within the jurisdiction of the Court' in Article 22(1) prevents the ICC from prosecuting crimes that were merely based on customary law but not defined in the Statute. 38 The emphasis on 'in accordance with this Statute' in Article 25(2) also indicates that the ICC is limited in its ability to hold an accused responsible merely based on a mode of liability that is recognised in customary law but goes beyond the scope of the Statute. 39 Considering these provisions of the Rome Statute, logically, customary law is becoming less significant than before at the ICC.
Customary law, however, still plays a role in the ICC framework. As noted above, customary law is considered as a secondary source; 40 in case of a legal gap, customary law may be referred to as an aid to interpreting the ICC's written rules. 41 The ICC has applied Article 21(1)(b) 42 and sometimes expressly confirmed customary law. In interpreting the qualifiers of 'widespread' and 'systematic' for the crimes against humanity, the Pre-Trial Chamber in the Katanga and Ngudjolo case cited ICTY and ICTR cases to confirm its interpretation of 'widespread or systematic'. 43 But in interpreting the term ' attack' as a war crime under Article 8 of the Rome Statute, the Trial Chamber in the Al Mahdi case did not consider the jurisprudence of the ICTY. In its view, these cases offer no guidance concerning the customary nature of the direction of attacks against civilian objects as a war crime. 44 The Ntaganda case deserves further attention. In clarifying a status requirement for victims of the war crime of rape and sexual slavery, the Trial Chamber in Ntaganda first concluded that the Rome Statute did not provide a status requirement and then turned to consider whether customary law requires such a limitation of status for victims. 45 The Prosecutor argued that introducing an additional element by relying on the expression ' established framework of international law' in Article 8 of the Statute would allow customary law to be applied even in the absence of a gap in the Statute. 46 To ensure consistency of Article 8 with international humanitarian law, the Appeals Chamber of the ICC rejected her argument and concluded that this expression allows the ICC to apply customary international law regardless of whether any lacuna exists. 47 In short, as for war crimes, an additional restrictive element in customary law may be applied by the ICC even if no legal gap exists.
In the drafting process of the Rome Statute, the ILC contemplated the inclusion of crimes under customary law. 48 In addition, according to Article 11(2), 13(b) and 24(1) of the Statute, the ICC may retroactively apply the Rome Statute to exercise jurisdiction over situations that occurred after its entry into force but before a State's ratification of the Statute or declaration of acceptance. In two contexts, individuals would be bound by customary law rather than the Statute. 49 Firstly, Article 12(3) of the Statute permits non-States Parties' retroactive acceptance of the ICC's jurisdiction by lodging a declaration with the Registrar. For example, through declarations in 2014 and 2015, Ukraine accepted the ICC's jurisdiction over alleged crimes committed from November 2013 onwards. 50 51 In the two contexts, how can the ICC retroactively exercise jurisdiction over these situations without violating the principle of legality prohibiting retroactive prosecution? As Broomhall wrote, '[t]he only legitimate basis for establishing the criminal responsibility of individuals [at the ICC] would presumably -in the absence of relevant national criminal prohibitions at the time of the alleged conduct -be that of customary international law'. 52 In the two contexts, customary law as a source of law does matter at the ICC. The ICC should clarify whether an offence in the Statute is a reflection of customary law at the material time.
The ICC was confronted with this circumstance in the Darfur Situation. In the Al Bashir case, Pre-Trial Chamber I tried to identify whether the rule in Article 27(2) reflects a new customary rule providing an exception to absolute personal immunity. 53 By contrast, the ICC missed the opportunity to do so in Côte d'Ivoire Situation. Given the fact that Côte d'Ivoire first accepted the jurisdiction of the ICC in 2003 and ratified the Rome Statute in 2013, it seems that the ICC should have engaged in discussing this issue when the Prosecution requested to broaden the scope of the investigation to cover alleged crimes committed from September 2002 onwards. Pre-Trial Chamber III directly applied the Rome Statute to expand the ICC's temporal jurisdiction without contemplating this issue. 54 Other situations and cases sharing the risk of retroactive application of the law reserve this issue for another day. In short, customary law continues to play a role in the ICC.
The examination above demonstrated that customary law is either a source or an interpretative aid of international criminal law. Customary law is and continues to be part of the applicable law of international and internationalised criminal tribunals. 55

III. Theories to Identify Customary International Law
This section examines academic theories to determine whether scholars have reached a consensus to create a different method of customary identification within international criminal law.

A. Identification of Customary International Law: the Classic Theory
In determining how a certain practice becomes a customary rule, the prevailing view is the presence of both subjective and objective elements. 56 Accordingly, the classic approach to identifying the existence of a customary rule is to seek sufficient evidence of these two elements, this is known as the two-element approach. State practice and opinio juris (opinio juris sive necessitates, the belief of law or necessity) are used mostly as the substitute of the objective element and the subjective element, respectively.
A large number of international scholars endorsed this two-element approach. 57 The Restatment of Foreign Relations Law of the United States also supported this idea. 58  Red Cross has employed this approach in its study on Customary International Humanitarian Law. 59 In 2012, the ILC included the topic 'Formation and Evidence of Customary International Law' in its agenda and appointed Sir Michael Wood as Special Rapporteur for this topic. 60 The title of this topic was later changed into 'Identification of Customary International Law'. Wood submitted five reports to the ILC with proposed conclusions. 61 Except for the use of the terms ' a general practice' and ' accepted as law' instead of 'State practice' and 'opinio juris', Wood also proposed the two-element approach. The Commission members welcomed this approach, and in 2018 the ILC adopted a set of 16 draft conclusions on 'Identification of Customary International Law'. 62 Its conclusion 2 under the title of 'two constituent elements' clearly reads that '[t]o determine the existence and content of a rule of customary international law, it is necessary to ascertain whether there is a general practice that is accepted as law (opinio juris)'. 63 It should be noted that the two-element approach remains the object of controversies among scholars supporting it. 64 This is in part because no consensus exists among international scholars as to what opinio juris consists of. Under the belief-theory, opinio juris is the belief of a State that the practice in question is a binding rule of international law. 65 This view is highly criticised for its circularity. Under the acceptancetheory, opinio juris means individual consent of a State. A State recognises that all States have a legal right to act in accordance with the practice. 66 This idea has been the dominant theory until now. Recently, Dahlman claims that the strong acceptance-theory should not stand, because each individual State's consent conflicts with the reality of international law. 67 In the alternative, he argues for a weak acceptance-theory that a large number of States approved a practice for all States to be bound as customary international law. 68 In his view, opinio juris functions as a filter preventing ' an unwanted general practice from becoming customary law'. 69 In addition to the divergence regarding the subjective element, there is also no strict standard that determines the weight of evidence required of each of the two elements. If the weight of evidence of one element is too slight, can we say that a rule would not be established and identified by relying on sufficient evidence of the other element? In 1950, Hudson put up criteria for the two elements but did not touch upon this issue in his working paper. 70 Kirgis' sliding-scale idea gives an illustration of this issue. 71 In his opinion, the evolution of customary international law must be examined on a sliding scale: one end is opinio juris, and the other end is State practice. The formation of a customary rule should be analysed on a case-by-case basis depending on different rules and acts. 72 This sliding-scale approach indicates that it is not possible for State practice to be zero percent and opinio juris to be one hundred percent, and vice versa. According to the idea of this sliding scale, more attention is paid to evidence of opinio juris than State practice for a moral-oriented rule. 73 Roberts also argued that State practice is becoming less important for rules with moral content. 74 Wood, in his second report to the ILC, confirmed that evidence of each element might be given different weight depending on the ' contexts'. 75

B. Identification of Customary Law in International Criminal Law: Another Theory?
As Schlütter observed, considering the high moral character of norms derived from international humanitarian law, such as the ' elementary consideration of humanity' contained in the Martens Clause, some scholars propose other approaches for the custom-formation and custom-identification. 76 Some commentators have proposed an opinio juris-based one element approach. Meron suggested the ' core right' approach for the custom-formation in international humanitarian law and international human rights law. 77 In his view, the content of customary law in the two fields would be inferred from the ' core values' of the international community. He also noted that 'Whether a violation of international criminal law triggered a broad condemnation by the international community is important for the establishment of customary norm'. 78 Some authors in international criminal law, for example, Pocar and Cassese, also support this approach if the rules belong to the 'canon of norms which can be held to represent the "core values" of the international community'. 79 It is true that in the field of international criminal law, the record of national investigations and prosecutions of international crimes is less satisfying. Evidence of State practice is more rarely obtainable compared to evidence of opinio juris. Nevertheless, it would be going too far to adopt this one-element ' core right' approach because it leaves room for powerful States to manipulate the law. 80 The ' core right' approach might also conflict with the strict principle of legality prohibiting analogy. Hart contended that international law differs from morality, and that the issues of international law are not about the moral issues of right or wrong and good or evil. 81 Koskenniemi also claims that ' elementary considerations of humanity' 82 do not fit into customary international law. 83 As Koskenniemi wrote: 'The social conception of law is in a dilemma: it cannot construe the normative sense of past behaviour in a bilateral relationship on the parties' real, psychological intent because such intent can neither be known nor authoritatively opposed to the State's own deviating view thereof. But it cannot base it on a non-psychological principle, either, because such principle will immediately look like a natural principle, based on non-verifiable and contested value preferences '. 84 This implies that the question of what a rule should be, reflecting the demanding values of protecting victims and prosecuting perpetrators, does not equate with the question of what a customary rule is in international criminal law. The identification of customary rules in international criminal law should not deviate from the two-element approach.

C. Summary and Observations
This section has shown that the majority of commentators supported the classic approach in general, although divergences exist among scholars regarding the content of the subjective element, that is opinio juris. Legal writers have proposed a flexible formulation of the two-element approach as to the weight of 75  evidence of each of the two elements. Until now, the one-element approach of ' core right' has not been widely accepted in academia. It seems that the classic two-element approach still applies in the identification of customary rules in international criminal law.

IV. Identification of Customary Rules of International Criminal Law by International Criminal Tribunals
This section endeavours to observe the identification approach employed by international and internationalised criminal tribunals. For this purpose, this section examines cases of these tribunals to discuss how the tribunals identified the existence of customary law in general and how they assessed the evidence of the two elements. It appears that the tribunals formally adopt the two-element method, while they reclassify evidence of the two elements in a flexible way. 85

A. The Two-Element Method: Formal 'Confirmation' or 'Silence'
It appears that international criminal tribunals have not expressly abandoned the two-element approach.
In the well-known Tadić decision on jurisdiction, the ICTY Appeals Chamber argued that '[w]hat is inhumane, and consequently proscribed, in international wars, cannot but be inhumane and inadmissible in civil strife'. 86 This decision has been criticised for its humanity-based reasoning. 87 91 The Appellant argued that there was no national legislation or military code as evidence of State practice. 92 The Appeals Chamber took this argument as conclusive. The Chamber found that 'no practice can be found, nor is there any evidence of opinio juris that would sustain' the customary international law that a superior is responsible for crimes committed before he or she assumed command over the subordinate. 93  The ICC has expressed the same view. 97 In the Malawi Cooperation Decision of the Al Bashir case, Pre-Trial Chamber I analysed whether sitting heads of State enjoy immunity in proceedings before international courts in international law. The Chamber held that 'initiating international prosecutions against Heads of State have gained widespread recognition as accepted practice', and that 'international community's commitment to rejecting immunity in circumstances […] has reached a critical mass'. 98 It then concluded that ' customary international law creates an exception to Head of State immunity when international courts seek a Head of State's arrest for the commission of international crimes'. 99 The Pre-Trial Chamber also formally endorsed the two-element approach to reach its conclusion. 100 In some cases, the wording of the two elements has been changed. When this occurs, the element of opinio juris tends to be an unnecessary requirement. 101 In Hadžihasanović et al., the ICTY Trial Chamber analysed whether there was a binding obligation on States to prosecute individuals for war crimes solely under international criminal law. 102 The Chamber found that there is no written rule for this issue. With regard to customary international law, it held that: 'To prove the existence of a customary rule, the two constituent elements of the custom must be established, namely, the existence of sufficiently consistent practices (material element), and the conviction of States that they are bound by this uncodified practice, as they are by a rule of positive law (mental element)'. 103 Based on national cases, the Chamber concluded that there was no consistent State practice. By inferring from the absence of sufficient consistent State practice, the Chamber found States were not obliged to prosecute war crimes under customary law at that time. 104 By referring to 'material element' and 'mental element' instead of 'State practice' and 'opinio juris', this Chamber adopted the two-element approach in general. However, the Chamber did not count opinio juris as an independent element for its inference that no opinio juris existed based on an absence of sufficient practice.
Secondly, several tribunals did not refer to the identification approach. The ICC, for example, was sometimes silent on the identification approach. In Ntaganda, the ICC determined whether it has jurisdiction over rape and sexual slavery as war crimes under Article 8(2)(e)(vi) of the Rome Statute when the alleged perpetrators belong to the same armed force as the victims. 105 The Appeals Chamber analysed whether customary law requires an additional element of the given war crime within the context of internal armed conflict, which limits victims of these two offences to 'persons taking no active part in hostilities' or 'persons who do not take part in hostilities'. 106 The Appeals Chamber rejected the existence of a customary rule requiring this status requirement for war crimes, either in general or in specific for the offences, without addressing its custom identification approach. 107 Other tribunals also failed to mention the two elements but relied on findings of the ICJ to confirm the existence of customary law. 108 By directly referring to the ICJ's finding in the Genocide Convention Reservation Advisory Opinion, the two ad hoc tribunals held that the 1948 Genocide Convention on the Prevention and Punishment of the Crime of Genocide (Genocide Convention) reflects customary law. 109 The ad hoc tribunals 97 Al Bashir Malawi Cooperation Decision (n 53) para 39. 98 ibid. paras 39, 42. 99 ibid. para 43. This finding, however, was rejected by Pre-Trial Chamber II of the ICC in the South Africa Cooperation Decision.
Without a further analysis, Pre-Trial Chamber II concluded that it 'is unable to identify a rule in customary international law that would exclude immunity for heads of State when their arrest is sought for international crimes by another State, even when the arrest is sought on behalf of an international court, including, specifically, this reached this conclusion based on whether other tribunals had recognised these treaty provisions as reflections of customary law. 110 This method of making direct references to ICJ decisions is very convenient and might be a way to assure coherency among international tribunals. The ICJ jurisprudence, which has no binding force ' except between the parties and in respect of that particular case', 111 might contain indications of custom. 112 The ICTY, however, seems to elevate these judicial decisions to the conclusive finding of the existence of custom. As for issues of international criminal law, international criminal tribunals should be more cautious about heavy reliance upon judgments of the ICJ.

B. Evidence of Two Elements
Despite using alternative methods at times, international criminal tribunals still try to identify customary rules by seeking sufficient proof of the two elements. These tribunals, however, have avoided coherently categorising materials as evidence of State practice or opinio juris. For instance, in Jelisic, the ICTY Trial Chamber tried to examine the element of genocide in customary law. 113  Even if tribunals have classified evidence in one of the two elements, the classification is not consistent. As for whether the existence of a treaty is evidence of customary international law, it has been classified as either opinio juris or State practice. The ICTY once held that the Rome Statute is evidence of customary law. 120  The Appeals Chamber in Tadić followed this idea. In Tadic, the Appeals Chamber relied specifically upon Article 25(3)(d) of the Rome Statute to establish a customary rule of the joint criminal enterprise. 123 It held that: 'The Statute was adopted by an overwhelming majority of the States attending the Rome Conference and was substantially endorsed by the Sixth Committee of the UN General Assembly. This shows that that text is supported by a great number of States and may be taken to express the legal position, i.e. opinio juris of those States'. 124 By contrast, in Krnojelac, the ICTY Appeals Chamber, with a reference to the Tadić Appeal Judgment, held that the Rome Statute is to be seen as evidence of State practice rather than opinio juris. 125 The Chamber stated that this ' analysis is also supported by recent State practice, as reflected in the Rome Statute'. 126 The SCSL followed the Krnojelac case in classifying the Rome Statute as a form of the two elements. In the Norman case, the SCSL examined whether the prohibition of recruiting child soldiers under the age of 15 existed as a war crime under customary law before November 1996. 127 To make this determination, the Appeals Chamber referred to the deliberations of the Rome Statute and proposals of States at the 1998 Rome Conference, and it deemed these statements 'State practice'. 128 The observation above shows that case law in international tribunals is not in a consistent fashion in respect of treaties as evidence of the two elements. The SCSL seems to differentiate State proposals concerning a treaty provision, as evidece of State practice, from the text of a treaty, as evidence of opinio juris.
Some Chambers even established the nature of a customary rule by simply referring to few exceptional interpretations. In Halilovic, the ICTY Trial Chamber examined whether a commander is responsible for the crimes committed by subordinates or for the failure to act under Article 7(3) of the ICTY Statute. 129 This Chamber referred to almost exactly the same materials as those in Čelebići. It firstly found that the post-World War II case law diverges about that issue. 130 Secondly, Additional Protocols to the 1949 Geneva Conventions did not clearly clarify the nature of command responsibility. 131 Finally, after examining the jurisprudence of the ICTY, the Chamber held that while most ICTY Chambers, including the Trial Chamber in Čelebići, had determined that a commander is responsible for the crimes committed by his or her subordinates under Article 7(3) of the Statute, there were, however, also a few exceptional interpretations. 132 The Trial Chamber concluded that commanders are responsible for the failure to prevent or punish crimes committed by their subordinates, which is the nature of command responsibility under customary law. 133 It should be noticed that the exceptions referred to in Halilović are the Aleksovski judgment and the partially dissenting opinion of Judge Shahabuddeen in the Hadžihasanović et al. Interlocutory Appeal decision. 134 The Krnojelac Appeal judgment seems to be another exception, while it was not referred to in Halilovic. 135 Turning to those cases however, shows that the Chambers in Aleksovski and Krnojelac did not consider the nature of command responsibility. In the Hadžihasanović et al. Interlocutory Appeal decision, Judge Shahabuddeen, in his partially dissenting opinion, also did not aim to clarify this issue. Instead, he provided an alternative interpretation of command responsibility as imposing the duty to punish, thus, a subsequent commander is liable for failing to punish crimes committed before his or her command assumption. 136 Therefore, the real exception in Halilović is the Halilović itself. Henceforth, ignoring the inconsistency of practice and unfavourable jurisprudence, the Chamber of Halilović relied on exceptional interpretations of command responsibility to conclude the nature of command responsibility under customary law. 137 It might be true that tribunals prior to Halilović made mistakes in the clarification of the nature of command responsibility under customary law. If this is the case, then what appears to be the exception, Halilovic, could be seen as a turning point and would be a valuable precedent for the development of customary law. Nevertheless, the Halilović Chamber neglected to identify inconsistent practice and adverse opinio juris, and reached a conclusion based solely on exceptional interpretations. Its reasoning is, therefore, less strongly persuasive. 138 The interpretative approach seems to avoid the two-element approach in the identification of customary law. It is unlikely that only frequent references to exceptional interpretations, without further practice and opinio juris, could create or modify a customary rule.

C. Summary and Observations
The observations of some cases from the international criminal tribunals indicate that there is a discrepancy between the words and deeds of tribunals in the identification of customary rules. Tribunals state that they will apply the two-element approach, while they do not distinguish evidence of State practice from that of opinio juris nor do they make consistent classifications of such evidence. Further, they sometimes do not investigate evidence of State practice or make a decision solely based upon opinio juris. The international tribunals have even established a customary rule by merely relying on few exceptional interpretations. Although tribunals do not always clarify the attributes of evidence, they give more weight to opinio juris than State practice in some cases. It seems that evidence of opinio juris is expanded and inclusive, while evidence of State practice is limited and exclusive.
To sum up, international criminal tribunals, in general, employ the two-element approach in a 'loose' or 'flexible' way to uphold or reject arguments regarding customary law. In fact, the two-element approach is also confirmed and acknowledged by other international bodies, for instance, the ICJ. 139 Its most outstanding example is the North Sea Continental Shelf cases. 140 The ICJ in the Libya/Malta case also stated that '[i]t is axiomatic that the material of customary international law is to be looked for primarily in the actual practice and opinio juris of states'. 141 The ICJ's approach has also been criticised by commentators because it is inconsistent and too flexible. 142 Commenting on the ICJ's case law, Judge Tomka, former President of the ICJ, has explained: In fact, the Court has never abandoned its view, firmly rooted in the wording of the Statute, that customary international law is 'general practice accepted as law' […] that is […] a settled practice together with 'opinio juris'. However, in practice, the Court has never found it necessary to undertake such an inquiry for every rule claimed to be customary in a particular case and instead has made use of the best and most expedient evidence available to determine whether a customary rule of this type exists. 143 Judge Tomka's remarks maybe also partly explain why these international criminal tribunals sometimes have not made a detailed analysis of the two elements and inconsistently classified evidence in specific cases.

V. Peculiarities of Identifying Customary Rules in International Criminal Law
As shown above, international criminal tribunals applied the two-element approach to identify customary law in a flexible way. Apart from Judge Tomka's 'best and most expedient evidence' explanation, the unique features of international criminal law and the difficulties in its customary identification are also persuasive explanations for flexibility in the application of the two-element method. Some commentators have also noted that ' customary international law in the context of international criminal law means something different from customary international law in the context of traditional international law'. 144 This section aims to explore some specific features for custom-ascertainment in international criminal law, including the prohibitory nature of international criminal law, the distinction between illegal and criminal behaviour, and the difficulties in the assessment of evidence.

A. Nature: Rules of Prohibition
In international criminal law, individuals must refrain from committing international crimes. National tribunals can investigate and prosecute suspected person for committing international crimes. However, not every State will prosecute international crimes for various political or legal reasons, for example, the lack of evidence or sources, the lack of motivation or scarcity of support in national law. Evidence of physical State practice, therefore, is more rarely obtainable.
Nevertheless, the fact that frequent violations exist and States have had few successful instances in practice does not impede the formation of a customary rule. 145 The prohibition of torture, for example, is recognised under customary law despite the frequent practice of torture around the world. 146 Some substantive rules of international criminal law derive their origins from international humanitarian law and international human rights law. 147 These two regimes introduce prohibitions such as the prohibition of torture and the abolition of death penalty. Multiple acts of international criminal law share the same characteristic, for example, the underlying offences of war crimes, crimes against humanity and genocide. Similarly, the fact that there is not a strong record of national investigation and prosecutions of international crimes is not an obstacle for the formation and the identification of customary law.

B. Illegality and Criminality of Behaviour
A distinction between unlawful behaviour and criminal behaviour must be kept in mind when identifying customary rules of international crimes. International law prohibits some conduct, and violations of its rules without justifications are regarded as an infringement. However, not all prohibitory norms entail individual criminal responsibility, for instance, not all violations of a rule of international humanitarian law constitute a war crime.
In Nahimana, the ICTR Trial Chamber examined hate speech as an underlying act of persecution under Article 3(h) of the ICTR Statute. Relying on this provision and Article 4 of the Convention on the Elimination of all Forms of Racial Discrimination (CERD) as well as Article 20 of the International Covenant on Civil and Political Rights (ICCPR), the Chamber held that 'hate speech expressing ethnic and other forms of discrimination violates the norm of customary international law prohibiting discrimination'. 148 The Trial Chamber concluded that Nahimana was guilty of persecution for his hate speech. In its view, hate speech merely advocating hatred constituted an underlying offence of persecution. 149 In the appeal, the Appeals Chamber found that hate speech alone as a violation to the right to respect for dignity constitutes an underlying discrimination. 150 According to the Appeals Chamber, 'it is not necessary that these underlying acts of persecution amount to crimes in international law'. 151 The Chamber, thererfore, did not decide whether mere hate speech is criminal under international law, but allowed it to be a basis for persecution. 152 In his dissenting opinion, Judge Meron held that mere hate speech is not criminal in customary law. 153 He clarified that: 'It is true that Article 4 of the CERD and Article 20 of the ICCPR require signatory states to prohibit certain forms of hate speech in their domestic laws, but do not criminalize hate speech in international law.
[…] In light of the reservations to the relevant provisions of the CERD and the ICCPR, the drafting history of the Genocide Convention, and the Kordić Trial Judgement, it is abundantly clear that there is no settled norm of customary international law that criminalizes hate speech '. 154 In his view, there was no customary law criminalising hate speech and mere hate speech cannot constitute an underlying act of persecution. The approach of the Appeals Chamber may violate the principle of legality. 155 The analysis of this case indicates that the distinction between illegal acts and criminalised acts should not be blurred. International law only criminalises serious acts rather than any violation of international human rights law as an international crime. The unlawful nature of violation does not lead to the criminalisation of it within international criminal law.

C. Obstacles in Evidence Assessment
Apart from individual State interests, a trend exists whereby the benefit of international community is becoming a focus of international criminal law. 156 Rules concerning the protection of human beings' common interest have progressively increased. Given customary rules in this area, there are some qualifications to assess evidence of practice, in particular, national cases and legislation.

National Cases and Legislation as Evidence of Practice
National cases and legislation are per se not sources of international law because most of them do not deal with international law issues. 157 However, as Oppenheim commented, national cases in ' cumulative effect' may afford evidence for the identification of customary law. 158 National cases and legislation addressing issues of international criminal law are relevant evidence of practice.
The first task in identifying relevant evidence of State practice is quite challenging. Given the complementarity principle of the ICC, the enforcement of international criminal law mostly relies upon States. 159 According to Article 20(3) of the Rome Statute, prosecution by the ICC is prohibited if the ' conduct proscribed' has been prosecuted by another court, thus, leaving the characterisation of ' crimes' open to national courts. 160 Therefore, the same conduct may be categorised as different crimes, and different conduct may be classified with the same label. One example is wilful killing as a war crime. Every State has universally recognised that wilful killing is a crime. On the one hand, wilful killing in armed conflict may be prosecuted and punished by national courts as ordinary crimes of murder, war crimes, or crimes against humanity at the international level. 161 For example, in 1973, a United States Army Lieutenant was convicted of murder and assault for his involvement in the My Lai massacre during the Vietnam War. 162 On the other hand, intentional killing in peacetime is prohibited and criminalised as murder in national law. In suppressing its political opponents, a government might also manipulate prosecutions and wrongfully prosecute some behaviour that constitutes ordinary crimes as war crimes. A Trial Chamber of the ICTY held that 'national prosecutions of war crimes characterised as ordinary crimes were not regarded as valid responses to interna-tional crimes in the ICTY Statute'. 163 In a nutshell, it is an arduous task to distinguish the irrelevant from the relevant evidence of practice supporting or rejecting an offence as an international crime.
Additionally, the scope and definition of international crimes in national law might not be the same as that under customary law. A good example would be the definition of terrorism in the Lebanese Criminal Code. The Appeals Chamber of the Special Tribunal for Lebanon (STL) 164 clarified that there is an emerging definition of terrorism as an international crime under customary law. 165 According to the Chamber, the elements of the crime of terrorism in the Lebanese Criminal Code are distinct from those under customary law. Customary law further requires the elements of an underlying act and the intent to commit that act. 166 The STL held that the elements of the notion of terrorism in Lebanese Criminal Code must be interpreted in accordance with international law by enshrining additional elements. 167 National courts indeed rarely examine the existence of a customary rule or try to interpret it in consonance with international law as the STL did. 168 In identifying evidence of practice, practitioners should be cautious about national decisions as evidence of customary law.
Moreover, some behaviour criminalised under international law is permitted in national law, while both rules run in parallel at the international and the national levels. For instance, the prohibition of expanding bullets in armed conflicts is recognised as a customary rule in international humanitarian law. 169 The use of expanding bullets in armed conflicts is also considered as a war crime under Article 8 of the Rome Statute. By contrast, the use of expanding bullets may be lawful in a law enforcement context at the national level. 170 The search on relevant evidence of national laws and their assessment becomes difficult because what is not justified in international law might be legitimate in national legislation. It is hard to seek sufficient evidence of national laws and cases to verify the existence of customary law. However, this idea does not mean that national cases do not matter.
National cases are sometimes hard evidence from which a customary rule is to be ascertained. It should be noted that the obligation to prosecute a crime ' as defined' under international law differs from the acknowledgement of that particular crime in international law. States may recognise international crimes in customary law, but they may be not obliged to prosecute underlying offences of these crimes solely ' as defined' in customary law. 171 As mentioned above, States may prosecute violations of war crimes as ordinary crimes. In this circumstance, these national prosecutions are valuable evidence for identifying the existence of international crimes in customary law.
The exercise of universal jurisdiction is another crucial piece of evidence for the identification of custom. In Filártiga v. Peña-Irala, a United States Circuit Court had to deal with a crime of torture that occurred in Paraguay. 172 In making the decision, the court examined whether the prohibition of torture had developed into a customary rule. 173 In Butare Four, Belgian courts brought four Rwandans to trial for war crimes. 174 This prosecution is seen as performing a gap-filling function with regard to jurisdiction because the ICTR declined to take charge of the prosecution and extradition of the accused back to Rwanda was impractical. 175 Similarly, in Scilingo, the Spanish Supreme Court concluded that unlawful detention and crimes of murder committed in Argentina amounted to crimes against humanity. 176 It is true that the concept of uni-versal jurisdiction and its requirements are controversial, and only a few national courts have successfully implemented universal jurisdiction. 177 These decisions exercising universal jurisdiction over international crimes, however, are significant evidence of State practice confirming a customary rule, especially when States concerned (Paraguay, Rwanda and Argentina) have also recognised such practice.

International Instruments as Evidence
International instruments are a form of evidence of opinio juris. 178 They should be given weight for the identification of customary rules in international criminal law. In using international instruments for this purpose, however, some words of caution are necessary. First, the definitions of crimes in international instruments may be broader or narrower than that in custom. The examples of war crimes and crimes against humanity in the ICTY and ICTR Statutes may suffice to illustrate this argument. The 1993 ICTY Statute stipulated a nexus with armed conflicts for the crimes against humanity. 179 The Appeals Chamber of the ICTY in Tadić argued however, that 'the requirement […] that crimes against humanity be committed in armed conflict, did not reflect customary international law, […] that the nexus with an armed conflict was not a new constitutive element of crimes against humanity'. 180 Additionally, in his report to the Security Council, the UN Secretary-General expressly noted that 'Article 4 of the Statute […] for the first time criminalizes common Article 3 of the four Geneva Conventions'. 181 This statement, at least, indicates his view that the scope of war crimes in Article 4 of the ICTR Statute was broader than the custom existing in 1994. Similarly, the ICTR Statute provides a discriminatory intent in the definition of crimes against humanity, whereas the ICTR argued that 'the Security Council decided to limit the jurisdiction of the Tribunal over crimes against humanity solely to cases where they were committed on discriminatory grounds'. 182 The ICTR concluded that the ' discriminatory intent' for crimes against humanity is not a substantive element. 183 Apart from adding new requirements, the scope of a crime in a treaty may be restricted by excluding some underlying acts; in particular, conventions that are not exhaustive codifications of existing customary law. 184 One example is the list of war crimes in non-international armed conflicts in Article 8 of the Rome Statute because it excludes some acts, such as launching an attack in the knowledge that it will cause excessive incidental civilian loss, injury or damage. Judge Schomburg of the ICTY also criticised the reference to Article 28 of the Rome Statute concerning command responsibility, as this treaty was a delicate compromise resulting from arduous negotiations. 185 The definition of genocide is another excellent example to illustrate this idea. The 1946 General Assembly Resolution 186 and the 1947 Draft Convention included political groups in their definitions of genocide, 187 although objections existed during the preparatory works as to the inclusion of such groups. States made compromises in order to adopt a treaty. The final 1948 Genocide Convention excludes the elimination of political groups from the definition of genocide. 188 Although many cases indicate that the 1948 Genocide Convention is reflective of customary law, there is a good possibility that the definition of genocide including political groups in the 1946 General Assembly Resolution mirrors the whole of customary law, whereas the 1948 Genocide Convention only codified part of customary law. 189 The drafters Tan 109 may have intentionally narrowed the definition of genocide. The example of the Genocide Convention also shows that preparatory works of international instruments are essential for the customary identification. Indeed, Governmental statements in international context, for example, notes, protests or claims, reactions to other States' claims, proposals or comments on drafts, are manifestations of practice.
Thirdly, it is necessary to note that in a treaty there is a difference between the obligations on States and the liabilities of individuals. In Akayesu, the Trial Chamber of the ICTR relied on the definition of torture in the UN Convention against Torture and Other Cruel, Inhuman and Degrading Treatment or Punishment (Torture Convention). 190 By contrast, in Kunarac, the Appeals Chamber of the ICTY held that the definition of torture in the Torture Convention is associated with obligations of States, which is not identical to the definition of torture as a crime against humanity. 191 The two cases show that it is necessary to identify whether a treaty, which aims to impose obligations on States, also attaches individual criminal responsibility for violations and provides clear definitions of international crimes.
Lastly, the retroactive application of international instruments should be avoided due to the risk of violating the principle of legality. Cases of the ad hoc Tribunals have confirmed this idea. 192 In Oric, Judge Liu doubted the reference to the Rome Statute because it was adopted after the ICTY and ICTR Statutes. As he points out, '[s]ince customary international law has to be assessed as of the date of commission of the offences, the fact that these texts were adopted subsequent to these dates, further limits their weight and usefulness as sources of customary international law at the time the crimes were committed'. 193

D. Summary and Observations
This section has shown that the evidence assessment involved in custom identification tends to be problematic because of the peculiarities of international criminal law and the obstacles in the evaluation of the evidence. These peculiarities mingling with each other make the identification task more complex, which should be kept in mind in determining the existence of customary law. These features partly serve to explain why international criminal tribunals practice flexibility when identifying customary rules in international criminal law. In reality, many States do not have capacity, or do not need, to prosecute international crimes in order to contribute to the formation of this custom. This thus, results in the paucity of States' physical acts. These tribunals give more attention to verbal statements of States and evidence of opinio juris than the States' physical acts. The evidence of opinio juris is raised to a higher importance in the identification of customary law. 194 As shown above, in section III, legal writers also accept a flexible formula of the two-element approach. Sørensen has suggested a practical way that 'in cases where a consistent practice can be proven, a certain presumption may arise in favour of the existence of opinio juris; so that the burden lies on the opposing party to […] refute the existence of a customary rule of law'. 195 Waldock and other scholars affirm this refutable "presumed" acceptance idea. 196 The converse of the 'presumed' acceptance idea also exists in international criminal law. It means that once sufficient opinio juris of a customary rule exists, less State practice is required to be enough for the identification of that rule. 197 The opposing party bears the burden to refute the presumed existence of a customary rule. A supposed customary rule and evidence of the two elements can be put on two sides of a scale. A customary rule is emerging, or modified, only if the side with the two-element acquired enough weight to either make the other side rise or balance the scale. For the change of a pre-existing customary rule, evidence of both elements is required to ascertain the content of a customary rule, while a denial of an existing rule would also be supported with sufficient opposing evidence of two elements.