‘To Destroy the Group Physically’? Part 1: The Genocide Convention, its Travaux, and Related Problems with the Dominant Interpretation

[Zsuzsanna Deen-Racsmány holds a Ph.D. in public international law from Leiden University. She is an independent researcher and has worked, inter alia, at Leiden University and the University of Amsterdam, and been a rapporteur for Oxford International Organizations.]

Introduction

Contrary to the earlier genocide cases before the International Court of Justice (ICJ or the Court) against Serbia (here and here), the nature of destruction that must be intended under the definition of the crime (in short: the scope of the dolus specialis) is not a prominent issue in the genocide cases currently on the Court’s docket. To be sure, in The Gambia v. Myanmar, the case closest to its conclusion, Myanmar hastened to emphasize the Court’s previous, to Myanmar favourable pronouncements on the issue (here, p. 34, para. 40 and here, para. 4.22), that the definition of genocide requires intent to destroy the group physically or biologically. However, The Gambia (p. 32, para. 26) did not dispute the validity of that now dominant reading of Article II of the Genocide Convention. The only cautious criticism came from an intervening state: Slovenia (para. 53) advocated for a broad interpretation of genocidal intent, albeit not specifically regarding the nature of destruction required. The ICJ has thus been given little reason to revisit its previous relevant holdings on that issue in its soon expected judgment.

This series of three posts, together with an earlier one, contend that the Court still has grounds to do so. In Part 1, I briefly argue that the terms of the Genocide Convention, its object and purpose, and its travaux préparatoires do not bear out the ICJ’s currently prevailing reading of the scope of the dolus specialis, presented in 2015 in Croatia v. Serbia. Part 2 examines the relevant holdings in the ICJ’s two existing judgments that appear to have addressed the subject. It demonstrates that certain Bosnia v. Serbia (2007) pronouncements have been wrongly perceived, even by the ICJ itself in Croatia v. Serbia, as having construed the dolus specialis narrowly: in Bosnia v. Serbia, the Court did not categorically find that genocide requires specific intent to destroy the group physically, nor would its reasoning corroborate that particular interpretation of the scope of the dolus specialis. In turn, in Part 3, I show that other relevant elements of Bosnia v. Serbia are consistent with this reading, and I explain why ICJ judges may even have been reluctant to express a categorical position on the scope of the dolus specialis in 2007. Finally, I argue that the Court did not and does not need to follow the (inaccurate) Croatia v. Bosnia holdings on that matter, which have come to be accepted as authority for the currently dominant interpretation.

The Definition of Genocide and the Scope of the Dolus Specialis

Article II of the Genocide Convention presents the two elements of genocide separately. The chapeau describes the dolus specialis, while the remainder of the provision specifies the actus reus. Whereas the concept of destruction is central to both, the direct object of destruction differs: the group as such—i.e., the social entity whose protection is a central goal of the Convention (Croatia v. Serbia, para. 139)—under the dolus specialis, but its individual members in the context of the actus reus. As explained by (then ICTY) Judge Shahabuddeen in his Partial Dissenting Opinion in Krstic (Appeal) (p. 102, para. 48), the nature of destruction pertaining to the dolus specialis and to the actus reus does not need to be identical. Indeed, the text of Article II is clear and does not contain such a requirement. Considering the markedly different nature of their respective protected objects, this is only logical: the ordinary meaning of the word ‘destroy’ differs in relation to groups and individuals, and—as the ICJ explicitly recognized—the “two constituent elements” of the crime of genocide are “analytically distinct” (Croatia v. Serbia, para. 130).

Among the five listed actus rei, the one stated in Article II(c) of the Genocide Convention (“[d]eliberately inflicting on the group conditions of life calculated to bring about its physical destruction”) is unique in that it contains a requirement similar to the dolus specialis. As noted by Quigley (p. 32), here, “proof of the [dolus specialis] seems to duplicate the intent required for the act itself”. However, that ‘duplication’ is not exact. Article II(c) covers only conduct calculated to result in the group’s ‘physical destruction’. Conversely, the chapeau of Article II makes no reference to physical (or biological) destruction in the context of the dolus specialis, and there is no evident reason for interpreting it as implying such a limitation either. The ordinary meaning of the words read in their context (including, a contrario, the explicit reference in Article II(c) to physical destruction) and consistently with the Convention’s object and purpose (to protect groups rather than individuals, emphasizing the value of their “cultural and other contributions” to humanity rather than any shared physical traits or genetic markers they may have) can well carry a broader reading, as even Serbia (p. 14, para. 12) acknowledged in the Croatia v. Serbia proceedings. The only factor commonly invoked in favour of the prevailing narrow interpretation of the dolus specialis is the Genocide Convention’s drafting history, in particular a decision (p. 206) of the Sixth Committee of the United Nations General Assembly to remove a provision from the penultimate draft of the Convention. 

That deleted provision—Article III of the Ad Hoc Committee draft (p. 55), described in a margin note as ‘cultural genocide’—specified two comparatively mild actus rei (the prohibition of the use of language and the destruction or restriction of the use of cultural institutions and monuments, libraries, schools, etc.) committed with intent to destroy the group’s culture, religion and language, i.e., its specific features rather than the group (the cohesive social unit) itself. Accordingly, it is difficult to perceive the contents of this draft article as describing genocide, either in terms of the dolus specialis or the actus reus. Although the earlier ‘cultural genocide’ draft provision, Article I(II)(3) of the Secretariat’s draft (pp. 6-7), included more violent forms of conduct and required intent to destroy the group, it too listed only means of destruction of culture, language and religion as actus reus. In other words, both drafts attached a very specific meaning to ‘cultural genocide’.

Nonetheless, the vote (p. 206) approving the proposal to remove draft Article III (popularly described as rejection of the ‘concept of cultural genocide’, but see Mundorff) has given rise to a much broader claim: that the Convention’s drafters thence decided to prohibit only conduct falling under the five considerably more heinous actus rei listed in paragraphs (a)-(e) of Article II of the Genocide Convention, committed with intent to destroy the group physically or biologically. Admittedly, records of the Convention’s travaux—in connection with this decision and the related Sixth Committee discussion (pp. 193-207), and as a whole—evidence several delegations’ (albeit not necessarily the majority’s) preoccupation with prohibiting physical-biological destruction. However, representatives appeared primarily concerned about the relative lack of gravity of conduct dealt with in draft Article III and the vagueness of that ultimately deleted provision. Even more importantly, speakers generally addressed factual destruction or the means of destruction, i.e., the actus reus. 

Conversely, the intended nature of the group’s destruction required in the context of the Convention’s ultimate dolus specialis was not discussed in any detail during the preparatory work, and a USSR proposal to insert ‘physical’ into the chapeau was promptly rejected (here pp. 95-97, cf. ibid., p. 96 for another relevant—but ignored—suggestion, by Peru). Accordingly, the records do not evidence a determination on the part of the majority of the Sixth Committee to require intent to destroy the group physically or biologically (see also Novic, ch. 3; Whitt and Clarke, ch. 7; other recent critical assessments of the travaux cited here; and Mundorff). 

After all, there is a considerable conceptual gap between the destruction of a group’s language, religion and culture on the one hand and its physical-biological destruction on the other, both in terms of intent and conduct. Most significantly, that gap encompasses conduct constituting any of the five actus rei committed with intent to destroy the group as a functioning social unit, more suitably described as social than as cultural genocide. According to its textual-contextual-teleological interpretation, Article II of the Convention prohibits also social genocide, and the records of the preparatory work—having not dealt with the nature of destruction that the dolus specialis must aim at in particular—do not contradict that reading. 

Significantly, whereas prominent genocide expert Schabas has apparently reached similar conclusions regarding the explicit contents of the travaux, he has nevertheless posited in relation to “the destruction that is part of the intent” that “the spirit of the discussions resists extending the concept of destruction beyond physical and biological acts” (pp. 233-234). Even if we accept that the rules of treaty interpretation as codified in the Vienna Convention on the Law of Treaties (VCLT) (Arts. 31-33) may justify recourse to the travaux—a supplementary means of interpretation—where the textual-contextual-teleological interpretation yields a clear and reasonable result, as in this case, Schabas’ oft-quoted observation appears to reverse the question, at least as it ought to be asked when interpreting the definition of genocide as a treaty provision. In my view, the point of departure ought to be a broad interpretation of the scope of the dolus specialis. After all, according to its textual-contextual-teleological interpretation following the general rule of treaty interpretation (VCLT, Art. 31), Article II of the Genocide Convention does not require intent to destroy the group physically or biologically. Moreover, Lemkin’s original conception of genocide (pp. 79-90) did not impose such a limitation either, and the cultural genocide provisions of the two official drafts of the Genocide Convention reflect his broad vision, which thus became part of the travaux at an early stage. Therefore, the relevant question before the ICJ appears to be whether the preparatory work—as such, rather than its spirit (see Novic, p. 81)—provides firm enough evidence to justify limiting the scope of the required intent, to physical and biological destruction only. 

As explained above, that question must be answered in the negative.

The Dominant Interpretation: What has Gone Wrong?

Mainstream international jurisprudence—starting with the interpretation of the scope of the dolus specialis in the 2001 Krstić judgment (paras. 571-580) of the Trial Chamber of the International Criminal Tribunal for the Former Yugoslavia, embraced by the ICJ, in any event in Croatia v. Serbia (paras. 136 and 510)—holds the opposite: that (genocidal) specific intent must aim at the group’s physical or biological destruction. However, on closer examination, the only sources cited in Krstić in favour of that narrow interpretation of the scope of the dolus specialis—the Genocide Convention’s preparatory work and a statement by the International Law Commission (ILC or Commission) that interpreted the definition of genocide in light of the Convention’s preparatory work—cannot corroborate that position. As discussed above, the Convention’s travaux did not specifically address the matter and do not confirm a conservative reading of the scope of the dolus specialis. In turn, the ILC’s statement (pp. 45-46, para. 12) is admittedly often invoked in support of that narrow interpretation. However, the Commission in fact presented it—after having discussed (specific) intent—in one of the paragraphs of its report where it considered the actus reus of genocide, and there are no indications that it nevertheless meant for it to pertain also to (specific) intent. Accordingly, the Trial Chamber appears to have failed to observe the distinction between the actus reus and the dolus specialis concerning the nature of destruction required, leading it to an incorrect interpretation of the scope of the latter.

In turn, Bosnia v. Serbia is commonly interpreted as having endorsed the Krstić position—albeit perhaps not “full-throated[ly]” (p. 230)—following that judgment (also) in conflating the actus reus with the dolus specialis. This reading of Bosnia v. Serbia is not implausible (in particular with regard to paragraph 344), even though Judge Shahabuddeen drew attention to that instrumental underlying flaw of the Krstić approach in his widely-noted partial dissent from the appeal judgment in that case already in 2004 (para. 48). However, as it will be shown in Part 2, the ICJ in Bosnia v. Serbia did not explicitly confirm the validity of the Krstić interpretation of the scope of the dolus specialis. In either case, the Bosnia v. Serbia pronouncements—on which the Court’s most recent holdings on that point presented in Croatia v. Serbia rest—are in fact not suitable to substantiate that currently dominant, narrow reading.