The gift that keeps on giving: The ICJ’s Orders on Provisional Measures in the Cases between Armenia and Azerbaijan by Massimo Lando
The gift that keeps on giving: The ICJ’s Orders on Provisional Measures in the Cases between Armenia and Azerbaijan
Assistant Professor, School of Law, City University of Hong Kong
Global Fellow, Centre for International Law, National University of Singapore
Two “tit-for-tat” requests for provisional measures
On 7 December 2021, the International Court of Justice (Court) handed down two orders in the parallel cases that Armenia and Azerbaijan brought against each other in September 2021. Both cases were brought under the 1965 International Convention on the Elimination of All Forms of Racial Discrimination (CERD) and originated from the 2020 armed conflict known as Second Nagorno Karabakh War in Armenia and Second Garabagh War in Azerbaijan. That conflict ended with a ceasefire brokered by the Russian Federation on 10 November 2020, but that did not stop the animosity between the two States, which eventually resulted in Armenia filing an Application against Azerbaijan on 16 September 2021, to which Azerbaijan responded by filing its own Application against Armenia on 23 September 2021. Both States claim that the other has breached its obligations under Articles 2–7 of CERD. Alongside their separate Applications, both States filed separate requests for provisional measures (Armenia’s request was part of its Application, for Azerbaijan’s request see here). Presumably because of the similarity of parties and claims, and the identical legal instrument involved, the Court heard the two requests for provisional measures back-to-back and handed down its orders on the same day. The discussion below focuses on the order in Armenia v. Azerbaijan while indicating relevant differences from that in Azerbaijan v. Armenia, which is in many respects identical to the former. Paragraph references thus are to the order in Armenia v. Azerbaijan, unless otherwise indicated.
The Court’s orders
Following its established approach, the Court first examined whether it had prima facie jurisdiction under Article 22 of CERD. To do so, the Court deconstructed prima facie jurisdiction into its two distinct components: prima facie existence of a dispute under CERD (paras. 19–29) and prima facie satisfaction of the procedural preconditions under Article 22 (paras. 30–42). As to the former, the Court considered the exchanges between the parties before the filing of the Applications to find that Armenia and Azerbaijan prima facie held diverging views concerning compliance with obligations arising under CERD (paras. 27–28). As to the latter, the Court recalled its finding, in Ukraine v. Russian Federation, that Article 22 established negotiation and proceedings before the CERD Committee as alternative procedural preconditions (para. 31); it then considered whether the parties had engaged in discussion with a view to resolving their dispute under CERD, finding that they had done so, at least on a prima facie level (paras. 39–41).
After its examination of prima facie jurisdiction, the Court turned to plausibility and the link between the rights invoked and the measures requested. Concerning plausibility, the Court stated that “[a] State party to CERD may invoke the rights set out in the above-mentioned articles only to the extent that the acts complained of constitute acts of racial discrimination as defined in Article 1 of [CERD]” (para. 58). With minimal explanation, the Court then simply held that “on the basis of the information presented to it by the Parties, […] at least some of the rights claimed by Azerbaijan are plausible rights under [CERD]” (para. 59). In Armenia v. Azerbaijan, the Court held that CERD did not plausibly require Azerbaijan to repatriate civilian detainees and prisoners of war and that “Armenia has not placed before the Court evidence indicating that these persons continue to be detained by reason of their national or ethnic origin” (para. 60). In Azerbaijan v. Armenia, the Court stated that CERD did not plausibly require Armenia to cease planting landmines or to enable Azerbaijan to undertake demining, and that “Azerbaijan has not placed before the Court evidence indicating that Armenia’s alleged conduct with respect to landmines has ‘the purpose or effect of nullifying or impairing the recognition, enjoyment or exercise, on an equal footing’, of rights of persons of Azerbaijani national or ethnic origin” (para. 53). On the link requirement, the Court found that a link existed between some of the rights invoked by Armenia and Azerbaijan and one of the measures requested (paras. 67–68).
The Court then examined the evidence of irreparable prejudice to the rights invoked by the parties, finding that this final requirement for indicating provisional measures was met in the circumstances (paras. 80–88). In Armenia v. Azerbaijan, the Court indicated that Azerbaijan had to: (i) protect persons captured in the 2020 conflict from violence and bodily harm and ensure their equality before the law; (ii) take all measures to prevent the incitement and promotion of racial hatred and discrimination; and (iii) take all measures to prevent acts of vandalism and desecration of Armenian cultural heritage (para. 98). In Azerbaijan v. Armenia, the Court indicated the same measures as (ii) above (para. 76). In both cases, the Court also indicated that the parties had to refrain from actions which might aggravate or extend the dispute.
Plausibility: the gift that keeps on giving (1)
From a procedural point of view, the Court’s approach to plausibility is the most interesting aspect of the two orders. The analysis in both Armenia v. Azerbaijan and Azerbaijan v. Armenia confirms the jurisprudence under which the Court assesses plausibility not only in respect of rights, i.e., whether the rights invoked might have a legal basis under international law, but also in respect of claims, i.e., whether evidence shows that respondents might have breached those invoked rights. This two-step inquiry into plausibility is different from the single-step examination of plausibility under the 2009 Belgium v. Senegal jurisprudence, which required the Court only to consider whether the rights invoked by an applicant State might have a basis under international law. This evolution of plausibility into a more demanding standard has already been the focus of scholarly discussion, with suggestions being made that plausibility evolved into such a standard in Equatorial Guinea v. France (see here) or in Ukraine v. Russian Federation (see here).
The cases between Armenia and Azerbaijan cement this evolution, but fail to provide appreciable clarity on the rationale for developing the plausibility standard to include a review of the evidence available at the provisional measures stage. At that stage, the Court is not supposed to look too closely into the evidence of the alleged breaches of international law yet, to avoid prejudging the merits of the case. However, one cannot deny that, since the introduction of plausibility in 2009, parties to provisional measures proceedings have frequently focused their submissions on matters of plausibility, often discussing questions of evidence at length. This tendency has brought about two consequences: first, provisional measures proceedings appear overburdened with references to evidence at a stage where such evidence is limited in amount and, thus, risks being misleading; second, the Court may feel compelled to address those references because the parties have pleaded them, which might result in prejudging the merits, at least to some extent.
The propensity to heighten the plausibility standard emerges also from the declaration of Judge ad hoc Keith. After having written that the humane treatment of persons arrested by State authorities and the preservation of cultural heritage are matters falling outside the material scope of CERD (plausibility of rights), he added that “Armenia has not provided […] evidence that these persons ‘continue to be detained by reason of their national or ethnic origin’” (plausibility of claims; see Declaration, paras. 4 and 8). Judge Yusuf made an identical comment (Dissenting Opinion, para. 8), but his dissent is notable especially because it shows the risk of conflating the analysis into plausibility with that into prima facie jurisdiction ratione materiae.
Commenting on plausibility, Judge Yusuf wrote that the acts of which Armenia complained “must plausibly constitute acts of racial discrimination […] and must be capable of falling within the scope of CERD” (Dissenting Opinion, para. 8). This statement conveys that, if the acts of which Armenia complained plausibly amounted to racial discrimination, they would also fall within the material scope of CERD. This conceptual overlap between plausibility and prima facie jurisdiction ratione materiae results from the similarity of the tests for both requirements: claims are plausible if the respondent’s conduct might breach the applicant’s rights; the Court has prima facie jurisdiction ratione materiae if the applicant’s claims prima facie “fall within the scope” of the treaty on which jurisdiction is founded, pursuant to the Oil Platforms test, which requires the Court to assess whether the conduct of the respondent might breach the rights that the applicant invokes under the relevant treaty. The similarity of test, which causes overlap in the assessment of plausibility and prima facie jurisdiction, means that the Court examines the same requirement under two different headings. Aside from a less than judicious use of time, this result leads to redundancy in the overall test for provisional measures.
One should ask: cui prodest? It is for the Court to give a satisfactory answer in future cases.
Another CERD case in a non-CERD case: the gift that keeps on giving (2)
The two cases between Armenia and Azerbaijan are only the latest of a growing number of disputes brought before the ICJ on the basis of CERD, the others being Georgia v. Russian Federation, Ukraine v. Russian Federation and Qatar v. UAE. The peculiarity of all these cases, including the ones discussed here, is that their underlying issues do not directly concern racial discrimination but, for instance, territorial sovereignty, international humanitarian law and restrictions on trade and flow of persons. Following the earlier cases, Armenia and Azerbaijan appear to be trying to fit their claims within the legal framework of CERD mainly to use the jurisdictional clause of that Convention to bring a case to the Court (for academic commentary on this phenomenon, see here).
In one of his Cambridge LL.M. lectures, the late Judge Crawford said that treaties sometimes are not important or useful because they create substantive rights for the States parties, but because they create jurisdiction for future disputes. CERD is one of those treaties. Over the past two decades, different States have brought cases based on the jurisdictional clause under CERD even though the disputes to which those cases related hardly concerned racial discrimination. On one level, one might fear for the future of participation in CERD: certain States might withdraw from the Convention if it becomes clear (as if it were not sufficiently clear already) that other States may frame claims as concerning racial discrimination only for the purpose to file ICJ cases which are, in truth, unrelated to racial discrimination. On another level, one wonders about the role of the Court as an agency for the peaceful settlement of inter-State disputes. It would be willing self-deceit to think that there are no political aspects to ICJ litigation. However, shoehorning claims into the legal framework of CERD to exert pressure on respondent States to comply with their non-CERD obligations risks transforming the judicial process before the Court into even more of a political instrument, which should not be the case in a (hopefully) increasingly mature international community.