Introduction
In its Order of 26 January 2024 in the case of South Africa v. Israel (brought under Article IX of the Convention on the Prevention and Punishment of the Crime of Genocide (“the Genocide Convention”)), the International Court of Justice (“ICJ”) indicated a number of provisional measures.
As part of its analysis, the Court found that “the facts and circumstances mentioned [in the Order] are sufficient to conclude that at least some of the rights claimed by South Africa and for which it is seeking protection are plausible” (¶54). The Court’s determination on “plausibility” appears to have given rise to significant confusion; at least in some of the commentaries analysing the Order, the Court’s finding on “plausibility” has been rephrased in a manner which departs from the way “plausibility” is traditionally understood in the Court’s jurisprudence. The apparent muddle caused by the Court’s reference to plausibility should not come as a complete surprise. Seemingly little has been written on the plausibility requirement and the threshold thereof, as well as on its exact application in the present case and its repercussions for the Court moving forward. This post seeks to broach precisely those questions.
What is the Plausibility Requirement?
In its 2009 Order on provisional measures in Belgium v. Senegal, the Court stated that its power “to indicate provisional measures should be exercised only if the Court is satisfied that the rights asserted by a party are at least plausible” (¶59), thereby introducing the plausibility requirement for the first time. More than 14 years later, in Guyana v. Venezuela, the Court described the requirement in substantially the same terms: “the Court may exercise this power only if it is satisfied that the rights asserted by the party requesting provisional measures are at least plausible” (¶19).
But what does plausibility of rights exactly mean? While the Court has never set out a comprehensive and clear explanation of this notion (e.g. Judge Nolte’s Declaration in South Africa v Israel, ¶10), a review of the Court’s jurisprudence on provisional measures can help shed further light on its meaning.
There are four relevant cases in which the ICJ declined to indicate provisional measures on grounds of failure to meet the plausibility requirement. It is possible to discern from these cases two main features in the Court’s determinations on ’plausibility of rights’:
Thus, in Qatar v. UAE (2019 Order), the Court declined indicating provisional measures solely on the ground that the claimed right did not plausibly exist (in the abstract, regardless of any factual assertion as to their violation) under the International Convention on the Elimination of All Forms of Racial Discrimination (CERD) – the treaty under which the Court had prima facie jurisdiction (¶¶ 25-26). Indeed, the UAE’s claimed rights related to procedural disadvantages in the ICJ proceedings, rather than substantive matters under CERD.
On one occasion, the Court declined to indicate provisional measures on the basis of the plausibility requirement only by reference to the facts of the case. In the case brought under the International Convention for the Suppression of the Financing of Terrorism (ICSFT) and CERD, Ukraine requested the Court to indicate several provisional measures aimed at ordering Russia to prevent terrorist financing. After observing that the ICSFT applies to financing only where there is intention or knowledge that funds will be used for terrorist acts, the Court observed that “Ukraine ha[d] not put before the Court evidence which affords a sufficient basis to find it plausible that these elements are present” (¶75). Indeed, Ukraine brought virtually no evidence purporting to point to intent or knowledge of Russian officials or nationals in their alleged financing of terrorist acts (CR 2017/3, pp 43-44).
On two further occasions, the Court declined to indicate provisional measures due to (lack of) plausibility of the claimed rights in the abstract, while also referring to the facts of the case:
In its Order of 26 January 2024 in the case of South Africa v. Israel (brought under Article IX of the Convention on the Prevention and Punishment of the Crime of Genocide (“the Genocide Convention”)), the International Court of Justice (“ICJ”) indicated a number of provisional measures.
As part of its analysis, the Court found that “the facts and circumstances mentioned [in the Order] are sufficient to conclude that at least some of the rights claimed by South Africa and for which it is seeking protection are plausible” (¶54). The Court’s determination on “plausibility” appears to have given rise to significant confusion; at least in some of the commentaries analysing the Order, the Court’s finding on “plausibility” has been rephrased in a manner which departs from the way “plausibility” is traditionally understood in the Court’s jurisprudence. The apparent muddle caused by the Court’s reference to plausibility should not come as a complete surprise. Seemingly little has been written on the plausibility requirement and the threshold thereof, as well as on its exact application in the present case and its repercussions for the Court moving forward. This post seeks to broach precisely those questions.
What is the Plausibility Requirement?
In its 2009 Order on provisional measures in Belgium v. Senegal, the Court stated that its power “to indicate provisional measures should be exercised only if the Court is satisfied that the rights asserted by a party are at least plausible” (¶59), thereby introducing the plausibility requirement for the first time. More than 14 years later, in Guyana v. Venezuela, the Court described the requirement in substantially the same terms: “the Court may exercise this power only if it is satisfied that the rights asserted by the party requesting provisional measures are at least plausible” (¶19).
But what does plausibility of rights exactly mean? While the Court has never set out a comprehensive and clear explanation of this notion (e.g. Judge Nolte’s Declaration in South Africa v Israel, ¶10), a review of the Court’s jurisprudence on provisional measures can help shed further light on its meaning.
There are four relevant cases in which the ICJ declined to indicate provisional measures on grounds of failure to meet the plausibility requirement. It is possible to discern from these cases two main features in the Court’s determinations on ’plausibility of rights’:
- whether it is plausible to read the treaty the applicant relies upon, as encompassing the rights alleged by the applicant in the abstract (i.e. regardless of the facts of the specific case); and
- whether any alleged facts were put forward by the applicant to support its argument as to violation of the rights in question.
Thus, in Qatar v. UAE (2019 Order), the Court declined indicating provisional measures solely on the ground that the claimed right did not plausibly exist (in the abstract, regardless of any factual assertion as to their violation) under the International Convention on the Elimination of All Forms of Racial Discrimination (CERD) – the treaty under which the Court had prima facie jurisdiction (¶¶ 25-26). Indeed, the UAE’s claimed rights related to procedural disadvantages in the ICJ proceedings, rather than substantive matters under CERD.
On one occasion, the Court declined to indicate provisional measures on the basis of the plausibility requirement only by reference to the facts of the case. In the case brought under the International Convention for the Suppression of the Financing of Terrorism (ICSFT) and CERD, Ukraine requested the Court to indicate several provisional measures aimed at ordering Russia to prevent terrorist financing. After observing that the ICSFT applies to financing only where there is intention or knowledge that funds will be used for terrorist acts, the Court observed that “Ukraine ha[d] not put before the Court evidence which affords a sufficient basis to find it plausible that these elements are present” (¶75). Indeed, Ukraine brought virtually no evidence purporting to point to intent or knowledge of Russian officials or nationals in their alleged financing of terrorist acts (CR 2017/3, pp 43-44).
On two further occasions, the Court declined to indicate provisional measures due to (lack of) plausibility of the claimed rights in the abstract, while also referring to the facts of the case:
- Armenia v. Azerbaijan (2021 Order): Armenia requested the Court to indicate provisional measures ordering the release of Armenian prisoners of war (POWs) detained by Azerbaijan during the 2020 hostilities between the two states. The ICJ noted that release of POWs was not plausibly governed by the CERD. It also observed that “Armenia ha[d] not placed before the Court evidence indicating that these persons continue to be detained by reason of their national or ethnic origin” (¶60; e.g. CR 2021/22, pp 21-23).
- Azerbaijan v. Armenia (2021 Order): Azerbaijan requested the Court to indicate provisional measures ordering Armenia to take all necessary measures to enable the former to demine landmines laid by Armenia and for Armenia to cease and desist laying mines on Azerbaijani territory. The ICJ noted that CERD does not plausibly govern acts relating to landmines. It observed that “Azerbaijan ha[d] 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” (¶53; e.g. CR 2021/26, pp 12-18).
To further zero-in on the threshold, it is also useful to consider instances where the ICJ found rights to be plausible, only to later reject the claim based on such alleged rights, either on the basis of the absence of jurisdiction ratione materiae or on the merits. In fact, since 2020, three cases in which provisional measures were in force have culminated. In one case – Equatorial Guinea v. France – the right found to plausibly exist at the provisional measures stage (diplomatic inviolability of the premises at 42 avenue Foch, Paris) was rejected at the merits stage by 9 votes to 7. Additionally, Equatorial Guinea’s claim based on that right was rejected by 12 votes to 4 (Judges Yusuf, Gaja and Sebutinde, who voted against the Court’s legal finding on the lack of diplomatic inviolability, nevertheless found that France’s actions did not constitute a violation in this case). In the second case – Qatar v. UAE – the Court found that it did not have jurisdiction ratione materiae by 11 votes to 6, thereby also rejecting (in the abstract) the existence of the alleged rights it previously found were plausible. In the third case – Ukraine v. Russia (ICSFT/CERD) – of the two provisional measures indicated relating to substantive rights found to be plausible, only one claim was upheld on the merits, while the other was rejected by 10 votes to 5.
To these cases, we may add Ukraine v. Russia (Genocide Convention). In its provisional measures Order, the Court found that “Ukraine has a plausible right not to be subjected to military operations by the Russian Federation for the purpose of preventing and punishing an alleged genocide in the territory of Ukraine” (¶60). By contrast, in its judgment on preliminary objections, the Court found by 12 votes to 4 that it lacked jurisdiction ratione materiae to adjudicate the part of Ukraine’s submissions which corresponded to the provisional measures Order, stating that such matters “are not governed by the Genocide Convention” (¶146).
Taking Stock
The ICJ’s jurisprudence suggests that the threshold of plausibility is very low, even nominal. One may question whether it even adds much to the already existing requirement of prima facie jurisdiction (compare Lando; Legality of the Use of Force). Indeed, it appears to differ distinctively from the type of examination usually applicable in both Common Law and Civil Law legal systems to issue provisional measures, which requires “at least some possibility of victory if the matter proceeds to final judgment”. Dissimilarly, the ICJ does not examine the strength of the claim made and its chances of success. Rather first, it merely considers whether there is a plausible link between the claimed rights and the relevant treaty; and second, that alleged facts were brought before the Court with regards to the claimed violation of those rights, regardless of their veracity or strength. Thus, in the few cases where the Court did consider the factual allegations it only rejected requests to issue provisional measures where the party requesting such measures brought no purported factual claims for the suggestion that the alleged rights could be in play in the circumstances of the case.
The unimpressive track record of rights which had been found to be plausible at the provisional measures stage, but which failed to be upheld at subsequent stages of the proceedings, is further testament to this low threshold. The relatively comfortable majorities by which previously found plausible rights were not upheld, or claims based thereon were rejected, suggests that the legal and evidentiary threshold for issuing provisional measures is almost symbolic. Arguably, the legal and evidentiary threshold is too low and should be reconsidered, but that normative assessment of the Court’s jurisprudence is not the focus of this post and will need to be treated separately.
Application in South Africa v. Israel
As noted in the introduction, the Court found that “the facts and circumstances mentioned above are sufficient to conclude that at least some of the rights claimed by South Africa and for which it is seeking protection are plausible”. Consistent with the Court’s jurisprudence, this finding simply means that (i) the rights claimed by South Africa’s plausibly existed under the Genocide Convention (in the abstract), and (ii) that this is not a case in which South Africa made no factual assertions aimed at supporting its claim that those rights have been violated. This finding does not mean that the Court assessed or opined on the strength of South Africa’s factual assertions. Indeed, in noting that “at least some” of the rights claimed by South Africa were plausible without elaborating, the Court implied that some of South Africa’s claims might not have been even linked to the Convention in a sufficient manner (¶54); and the Court further added its standard reiteration that its determinations in no way prejudge future findings on jurisdiction, admissibility or the merits of the case (¶¶15, 30, 84). The Court maintained the “plausible rights” terminology in its recent Order of 28 March 2024 (¶¶ 26, 27, 40).
Certain scholars rephrased the Court’s determination on plausibility by saying that “it is plausible that there is a case under the Genocide Convention” (here); “that South Africa’s claim of a violation of rights under the Genocide Convention is ‘plausible’” (here); “that South Africa made a plausible case that Israel is violating the Genocide Convention” (here); or even that “it was plausible that genocide was being committed in Gaza” (here, emphasis in original). Some media publications that described the Court’s decision followed through. However, such reformulations are neither reflected in what the Court actually said, nor – as the analysis in this post demonstrates – would they conform to the Court’s pattern of jurisprudence since first coining “plausibility” in 2009. The Court’s original language and intent were so broadly misrepresented, that the President of the Court at the time of the decision (now retired), Joan Donoghue, recently saw a need to stress in a BBC interview that contrary to publications, the Court did not decide that the claim of genocide was plausible, but rather, merely addressed the ‘plausibility of rights’.
As explained, the meaning of the Court’s finding on plausibility – in light of its previous case law – is that the Court found that South Africa invoked rights that are explicitly stipulated in the Genocide Convention, and that it presented factual arguments that purported to at least partially correspond with those rights. Indeed, based on the separate opinions and declarations issued by judges who had voted in favour of some of or all the measures, it is questionable whether South Africa was even required by the Court to raise factual arguments at all on all the Convention’s definitional elements necessary to establish a violation of its terms. For example, the declaration of Judge Bhandari (who voted with the majority on all the measures) reveals that to get his vote, South Africa merely had to indicate detrimental consequences to the civilian population in Gaza as a result of the war, without requiring any alleged evidence indicating genocidal intent on the part of Israel – although it is an essential condition for the commission of genocide under Article II of the Convention. Similarly, in his declaration appended to the additional Order issued on 28 March Judge Yusuf opined that “[t]here is no need for the Court at the stage of indication of provisional measures to determine the existence of genocidal intent” (a position somewhat difficult to reconcile with Judge Yusuf’s concerns repeatedly aired in Armenia v. Azerbaijan that “claims under humanitarian law have been given a home in CERD”). This approach seems to follow the Court’s decision in Gambia v. Myanmar, where intent was completely side-lined. In other words, at least in cases based on the Genocide Convention, it seems that even meeting the nominal evidentiary threshold set in Ukraine v. Russia (ICSFT/CERD) is not required. On this view, almost any armed conflict could be the basis for provisional measures by the Court.
In response to Judge Donoghue’s clarification on the South Africa v. Israel decision, some commentators tried to still argue that the Court found that South Africa’s claim of genocide was plausible based on another determination by the Court, that “there is a real and imminent risk that irreparable prejudice will be caused to the rights found by the Court to be plausible” (¶ 74) (here, here and here). However, apart from the fact that this text is not helpful in understanding what ‘plausibility’ means and does identify the specific rights that the Court found plausible, it reflects a forward-looking assessment rather than a determination that a violation may have occurred in the past. Indeed, if the Court’s remark on the risk for protected rights would have been tantamount to saying South Africa’s claims were plausible, presumably Judge Donoghue would not have felt a need to correct the misleading publications.
A Forum for Hit-and-Runs?
Scholars have been alert to the phenomenon of misuse of the provisional measures procedure for some time now. The ease in which a state can prevail in a request for provisional measures, even when it has little (or no) chance of prevailing on the merits is troubling.1 The South Africa v. Israel case demonstrates exactly this point, at least if one accepts (as I do) Marko Milanovic’s prediction before the Court’s Order in South Africa v. Israel that “[j]ust like I am virtually certain that South Africa will ‘win’ on provisional measures and get some kind of order, I am in the same way virtually certain that South Africa will lose on the merits.”
Indeed, South Africa’s weak evidence purporting to prove a violation of the Genocide Convention; its unprecedented 72-page Application – a document which should be limited to “a succinct statement of the facts” (Article 38(2), Rules of Court) – which reads more like a memorial, to set the stage for its request for provisional measures (compare with The Gambia v. Myanmar (15 pages)); its repeated requests to receive further provisional measures weeks after the Order (here and here), all seem to suggest that the success on provisional measures – rather on the merits – was, and remains, a key component of its strategy.
How this phenomenon affects state behaviour is still early to assess. Yet, if this trend continues and cases see provisional measures awarded, signalling conclusions which at the merits stage are found to be unwarranted, there is a real possibility of eroding the normative force of the ICJ’s provisional measures regime.
In any event, until the Court potentially changes course with regard to the threshold it applies for granting provisional measures, it would be advisable not to read any substantive conclusions into the mere issuance of provisional measures.
To these cases, we may add Ukraine v. Russia (Genocide Convention). In its provisional measures Order, the Court found that “Ukraine has a plausible right not to be subjected to military operations by the Russian Federation for the purpose of preventing and punishing an alleged genocide in the territory of Ukraine” (¶60). By contrast, in its judgment on preliminary objections, the Court found by 12 votes to 4 that it lacked jurisdiction ratione materiae to adjudicate the part of Ukraine’s submissions which corresponded to the provisional measures Order, stating that such matters “are not governed by the Genocide Convention” (¶146).
Taking Stock
The ICJ’s jurisprudence suggests that the threshold of plausibility is very low, even nominal. One may question whether it even adds much to the already existing requirement of prima facie jurisdiction (compare Lando; Legality of the Use of Force). Indeed, it appears to differ distinctively from the type of examination usually applicable in both Common Law and Civil Law legal systems to issue provisional measures, which requires “at least some possibility of victory if the matter proceeds to final judgment”. Dissimilarly, the ICJ does not examine the strength of the claim made and its chances of success. Rather first, it merely considers whether there is a plausible link between the claimed rights and the relevant treaty; and second, that alleged facts were brought before the Court with regards to the claimed violation of those rights, regardless of their veracity or strength. Thus, in the few cases where the Court did consider the factual allegations it only rejected requests to issue provisional measures where the party requesting such measures brought no purported factual claims for the suggestion that the alleged rights could be in play in the circumstances of the case.
The unimpressive track record of rights which had been found to be plausible at the provisional measures stage, but which failed to be upheld at subsequent stages of the proceedings, is further testament to this low threshold. The relatively comfortable majorities by which previously found plausible rights were not upheld, or claims based thereon were rejected, suggests that the legal and evidentiary threshold for issuing provisional measures is almost symbolic. Arguably, the legal and evidentiary threshold is too low and should be reconsidered, but that normative assessment of the Court’s jurisprudence is not the focus of this post and will need to be treated separately.
Application in South Africa v. Israel
As noted in the introduction, the Court found that “the facts and circumstances mentioned above are sufficient to conclude that at least some of the rights claimed by South Africa and for which it is seeking protection are plausible”. Consistent with the Court’s jurisprudence, this finding simply means that (i) the rights claimed by South Africa’s plausibly existed under the Genocide Convention (in the abstract), and (ii) that this is not a case in which South Africa made no factual assertions aimed at supporting its claim that those rights have been violated. This finding does not mean that the Court assessed or opined on the strength of South Africa’s factual assertions. Indeed, in noting that “at least some” of the rights claimed by South Africa were plausible without elaborating, the Court implied that some of South Africa’s claims might not have been even linked to the Convention in a sufficient manner (¶54); and the Court further added its standard reiteration that its determinations in no way prejudge future findings on jurisdiction, admissibility or the merits of the case (¶¶15, 30, 84). The Court maintained the “plausible rights” terminology in its recent Order of 28 March 2024 (¶¶ 26, 27, 40).
Certain scholars rephrased the Court’s determination on plausibility by saying that “it is plausible that there is a case under the Genocide Convention” (here); “that South Africa’s claim of a violation of rights under the Genocide Convention is ‘plausible’” (here); “that South Africa made a plausible case that Israel is violating the Genocide Convention” (here); or even that “it was plausible that genocide was being committed in Gaza” (here, emphasis in original). Some media publications that described the Court’s decision followed through. However, such reformulations are neither reflected in what the Court actually said, nor – as the analysis in this post demonstrates – would they conform to the Court’s pattern of jurisprudence since first coining “plausibility” in 2009. The Court’s original language and intent were so broadly misrepresented, that the President of the Court at the time of the decision (now retired), Joan Donoghue, recently saw a need to stress in a BBC interview that contrary to publications, the Court did not decide that the claim of genocide was plausible, but rather, merely addressed the ‘plausibility of rights’.
As explained, the meaning of the Court’s finding on plausibility – in light of its previous case law – is that the Court found that South Africa invoked rights that are explicitly stipulated in the Genocide Convention, and that it presented factual arguments that purported to at least partially correspond with those rights. Indeed, based on the separate opinions and declarations issued by judges who had voted in favour of some of or all the measures, it is questionable whether South Africa was even required by the Court to raise factual arguments at all on all the Convention’s definitional elements necessary to establish a violation of its terms. For example, the declaration of Judge Bhandari (who voted with the majority on all the measures) reveals that to get his vote, South Africa merely had to indicate detrimental consequences to the civilian population in Gaza as a result of the war, without requiring any alleged evidence indicating genocidal intent on the part of Israel – although it is an essential condition for the commission of genocide under Article II of the Convention. Similarly, in his declaration appended to the additional Order issued on 28 March Judge Yusuf opined that “[t]here is no need for the Court at the stage of indication of provisional measures to determine the existence of genocidal intent” (a position somewhat difficult to reconcile with Judge Yusuf’s concerns repeatedly aired in Armenia v. Azerbaijan that “claims under humanitarian law have been given a home in CERD”). This approach seems to follow the Court’s decision in Gambia v. Myanmar, where intent was completely side-lined. In other words, at least in cases based on the Genocide Convention, it seems that even meeting the nominal evidentiary threshold set in Ukraine v. Russia (ICSFT/CERD) is not required. On this view, almost any armed conflict could be the basis for provisional measures by the Court.
In response to Judge Donoghue’s clarification on the South Africa v. Israel decision, some commentators tried to still argue that the Court found that South Africa’s claim of genocide was plausible based on another determination by the Court, that “there is a real and imminent risk that irreparable prejudice will be caused to the rights found by the Court to be plausible” (¶ 74) (here, here and here). However, apart from the fact that this text is not helpful in understanding what ‘plausibility’ means and does identify the specific rights that the Court found plausible, it reflects a forward-looking assessment rather than a determination that a violation may have occurred in the past. Indeed, if the Court’s remark on the risk for protected rights would have been tantamount to saying South Africa’s claims were plausible, presumably Judge Donoghue would not have felt a need to correct the misleading publications.
A Forum for Hit-and-Runs?
Scholars have been alert to the phenomenon of misuse of the provisional measures procedure for some time now. The ease in which a state can prevail in a request for provisional measures, even when it has little (or no) chance of prevailing on the merits is troubling.1 The South Africa v. Israel case demonstrates exactly this point, at least if one accepts (as I do) Marko Milanovic’s prediction before the Court’s Order in South Africa v. Israel that “[j]ust like I am virtually certain that South Africa will ‘win’ on provisional measures and get some kind of order, I am in the same way virtually certain that South Africa will lose on the merits.”
Indeed, South Africa’s weak evidence purporting to prove a violation of the Genocide Convention; its unprecedented 72-page Application – a document which should be limited to “a succinct statement of the facts” (Article 38(2), Rules of Court) – which reads more like a memorial, to set the stage for its request for provisional measures (compare with The Gambia v. Myanmar (15 pages)); its repeated requests to receive further provisional measures weeks after the Order (here and here), all seem to suggest that the success on provisional measures – rather on the merits – was, and remains, a key component of its strategy.
How this phenomenon affects state behaviour is still early to assess. Yet, if this trend continues and cases see provisional measures awarded, signalling conclusions which at the merits stage are found to be unwarranted, there is a real possibility of eroding the normative force of the ICJ’s provisional measures regime.
In any event, until the Court potentially changes course with regard to the threshold it applies for granting provisional measures, it would be advisable not to read any substantive conclusions into the mere issuance of provisional measures.