Consensus, at what Cost?

After four applications for provisional measures, three sets of formal orders and two rounds of oral hearings, on Friday night, the International Court of Justice in South Africa v. Israel delivered a long-awaited Order. In paragraph 2(a) of the operative clause, the Court declares that Israel shall:

Immediately halt its military offensive, and any other action in the Rafah Governorate, which may inflict on the Palestinian group in Gaza conditions of life that could bring about its physical destruction in whole or in part;

This order was passed with 13 votes in favour and 2 against, an apparently strong majority. News outlets ran headlines shortly afterwards declaring that the Court had ordered Israel to “immediately halt its military assault on the southern Gaza city of Rafah”. The situation in Gaza is “catastrophic”, as the Court points out. Surely this is a landmark moment, one that will save lives on the ground. So why instead is everyone arguing about a comma, and why does it even matter?

Let’s Eat, Grandma or Let’s Eat Grandma: Punctuation Matters

There are multiple interpretations of this part of the Order doing the rounds, predominantly based on the placement of the commas and the use of the word “may”. Stefan Talmon contends that the sentence structure means that “Israel was not ordered to halt all military operations but only military operations that would violate its obligations under the Genocide Convention”. Adil Haque, acknowledging the Order is “somewhat ambiguous”, suggests that in practical terms it means the military offensive in Rafah “must immediately halt”. Heidi Matthews observes that the Court has ordered Israel to “immediately end all military activities in Rafah” and in doing so “has effectively held that the humanitarian disaster is such that it isn’t possible to conduct an offensive in a manner consistent with international humanitarian law”. Kevin Jon Heller and Mike Becker, among others, also argue that the Court’s order needs to be read in light of the Court’s factual findings (at paragraphs 45-47 in particular) that the offensive risks creating conditions of life that might lead to physical destruction of the Palestinian people, and therefore the ‘condition’ that follows the second comma is already met.

Even more intriguing, Judges have indicated in separate opinions that there may be different opinions about the interpretation of this clause between members of the Court. Judge Nolte, although in the majority, considers that the Court has ordered Israel to “limit the current military offensive in Rafah as far as it could endanger the rights of the Palestinian people under the Genocide Convention, notably their access to basic humanitarian needs”. Judge Aurescu makes the observation that:

I consider that the second provisional measure indicated … is somehow unclear as to whether the last part of it (starting with “which may inflict”) only refers to “any other action” (which is not defined) or to both halting the Israeli military offensive and “any other action”. In my view, this measure needs to be interpreted that it indicates as well the halt of the Israeli military offensive to the extent that it “may inflict on the Palestinian group in Gaza conditions of life that could bring about its physical destruction in whole or in part”.

Judge Sebutinde, in the minority, contends that the relevant paragraph “could be erroneously misunderstood as mandating a unilateral ceasefire in part of Gaza”. But then Judge Tladi tells us that the Court has “in explicit terms, ordered the State of Israel to halt its offensive in Rafah.”

This blog post is not going to offer yet another interpretation of the already infamous clause. Instead, it asks why and how the Court, faced with such a high profile and important decision, could end up delivering an Order which has left academics, journalists, lawyers and even the Judges themselves arguing over what it means.

Consensus versus Clarity: How the Court Decides

We must begin with a disclaimer: the author has no secret insight into what has gone on behind the scenes at the Peace Palace in The Hague. But if we reflect on the Court’s methods of decision making, it appears that in this instance the Court may have been driven by a desire to convince as many Judges as possible to vote in favour of the Order, at the cost of issuing a clearer and more straightforward directive. Quite possibly, the Court has deliberately adopted a phrasing which can be interpreted more than one way in order to get the decision across the line.

The Court’s official decision making procedure is set out in the Resolution Concerning the Internal Judicial Practice of the Court, most recently updated in 2023. The Court reaches a decision through a process of methodical joint deliberation. In ordinary proceedings, a deliberation is held at which the President outlines the issues that will require discussion and decision by the Court. Judges may comment on the pertinence of any issues or questions arising in the case and offer their views thereon. Judges then prepare their individual notes which set out their views on the questions which the Court should answer and a tentative conclusion as to the correct disposal of the case. In provisional measures cases the process is organised slightly differently due to time constraints, but it still creates opportunities for discussion and compromise, which are described by Tommaso Soave (Everyday Makers, p.281) as follows:

“A debate then ensues where the members of the Court seek to persuade each other of the merits of their opinions, resist countervailing arguments, make and unmake alliances to send their points across.”

The Registry staff may also be involved; preparing a statement of the facts or checking the draft orders for ambiguities (for a colourful autobiography of Hugh Thirlway’s activities in this respect during his tenure as Principal Legal Secretary, see here). Indeed, Thirlway admits that there were times during urgent provisional measures hearings that he presented the Judges tasked with drafting “a complete draft Order, prepared on the basis of the oral deliberation”. Ordinarily, the draft text goes through a first and second reading, with amendments to the text able to be proposed by the Judges (although again, for provisional measures this can be truncated). Finally, a formal vote is taken, in inverse order of seniority.

What does this procedure reveal to us? First, that the Judges are, from the outset, aware of the views of their fellow members of the Court. While the existence of any majority remains fluid until the final vote is rendered, it may only emerge once Judges are content with the formulation of the text. To put it in clear terms, it is entirely possible that certain Judges needed to be talked into agreeing with a ceasefire order, and the ambiguous paragraph at issue was the version everyone could agree upon.

This isn’t to imply any Judges were holding out for illegitimate reasons. The Court’s jurisdiction in this case is limited to the application of the Genocide Convention. There may well have been concerns that a ceasefire order needed to be linked closely to the Court’s jurisdiction to avoid overreach. More cynically, one could suggest that an ambiguous formulation leaves room for plausible deniability at the merits stage – but it’s not clear in which direction. Israel can argue that it has complied with the Order if it continues military operations in a way which does not inflict on the Palestinian group in Gaza conditions of life that could bring about its physical destruction in whole or in part. But equally, South Africa can argue that Israel has failed to comply with the Order if it continues its military operation in Rafah at all, especially in view of the reasons on which the Court based its determination in the Order that the new circumstances in Rafah demonstrated an urgent and serious risk or irreparable harm to the rights at issue in the case. Is the Court simply kicking the can down the road, waiting to resolve this question at the merits stage?

It is, to be frank, most unsatisfactory. While the Court is known for its “Solomonic” decisions, which try to give each party a little of what they asked for at times to no one’s satisfaction, this is not a maritime boundary delimitation where equidistance can be imposed in pursuit of impartiality. This Order is a demand, of Israel, to take certain concrete steps. It is unfair to Israel to be unclear in what is expected of it, and it is potentially ruinous for the people of Rafah should interpretation A be applied when interpretation B was intended.

As David Luban argues (in Tasioulas & Besson, The Philosophy of International Law, p.579), international courts “bootstrap themselves into legitimacy by the quality of justice they deliver”. The Court is indeed “only a court”, as Judge Tladi states, but that is also exactly what it is. And the role of a Court is to make decisions, sometimes very hard ones. It cannot split the baby. Consensus can be a useful thing, but a Court does not need consensus as much as it needs clarity. Better a clear and unambiguous Order, with fewer votes in favour. At least then, there would be no question as to what the parties are intended to do. As of May 25, the day after the decision, Israel continues to bomb Rafah. But we will not be certain that this is in defiance of the Court for some time to come. And all because of a comma.