(Un)Common Understandings

On 21 May 2026, the International Court of Justice (the Court) issued its opinion on the Right to Strike, for which the Court was concerned with the question of whether “the right to strike of workers and their organizations is protected under the Freedom of Association and Protection of the Right to Organise Convention, 1948 (No. 87) (Convention No. 87)” which it answered in the affirmative. This is a core matter of treaty interpretation and since the Court gave its opinion, several comments on its approach to treaty interpretation have been published (see here, here and here). We turn to a yet underexplored component of the opinion: The Court’s approach to Art. 31(3)(c) of the Vienna Convention on the Law of Treaties (VCLT) in the light of its engagement with State silence and the normative content of the applicable rules of international law.

The Court’s “Common Understanding”

The advisory opinion has its origin in a decade-long dispute within the International Labour Organization (ILO) since the Convention No. 87 does not mention the right to strike. Central to the Courts affirmation was its recourse to other treaties by way of Art. 31(3)(c) VCLT, which stipulates that other relevant rules of international law applicable in relations between the parties shall be taken into account when interpreting a treaty. It begins its analysis with a direct quotation of Art. 31(3)(c) VCLT and confirms that it reflects customary international law. Immediately afterwards, the Court turns to the main question: the threshold for the “applicable in the relations between the parties” component. The Court holds that not all parties to the treaty subject to interpretation must be “bound by the ‘relevant rules of international law’ in order for those rules to be taken into account” (para. 90). It suffices that the rule “expresses their common understanding regarding certain provisions of the treaty under interpretation” (para. 90). Such a common understanding may be inferred from a high degree of overlap in treaty membership.  Accordingly, where the parties to the treaty under interpretation have implicitly accepted rules in other treaties, these rules show a wide adoption, leading the Court to affirm the applicability (para. 92).

Why the Court requires this threshold becomes clear in para. 91: No other relevant treaty rule binding upon all States parties to Convention No. 87 exists. Nevertheless, the Court identifies two treaties of importance for the interpretation of the right to strike: The International Covenant on Civil and Political Rights (ICCPR) and the International Covenant on Economic, Social and Cultural Rights (ICESCR). In ascertaining whether the threshold is met, non-parties to the ICCPR and ICESCR and States with reservations in place are of particular interest (para. 92). In the present case, the non-party States are Comoros, Cuba, Kiribati and Saint Lucia. Whereas the former three States have recognized a right to strike in their domestic legal systems in the past, Saint Lucia has not done so. However, the Court reasons that because Saint Lucia never objected to the interpretations of the ILO’s supervisory bodies asserting a right to strike under Convention No. 87, it accepts that such a right is protected under Convention No. 87. Moreover, according to the Court, the relevant reservations mostly limit the scope of the right to strike, enshrined in Art. 8(1)(d) ICESCR, to certain categories of workers, without contesting the right to strike as such. Applying this standard, the Court relies on Arts. 8 ICESCR and 22 ICCPR, the latter containing the right to freedom of association and trade unions, as relevant rules of international law applicable between the parties and to that end cites both the Committee on Economic, Social and Cultural Rights (CESCR) and the Human Rights Council (HRC) (para. 96).

The Opinions

Judge Tomka opposes the Court’s methodological approach. He emphasizes the wording of Art. 31(3)(c) VCLT and, most notably, equates the term “the parties” with “all parties”. This signals a formalist-voluntaristic approach, aimed at including instances of express, formal consent to the requirements of systemic integration. He prefers this strict approach, as “the absence of clear criteria [would] render […] the identification of [such] a rule […] inherently speculative.” (Dis. Op. Tomka, para. 66). Consequently, such a strict approach would structurally limit the range of rules applicable under Art. 31(3)(c) VCLT. Judge Tomka further draws on the International Law Commission’s (ILC) work on the law of treaties to demonstrate that the meaning of Art. 31(3)(c) VCLT was originally understood to require the relevant rule to be binding on all parties to the treaty under interpretation. Similarly, Judge Hmoud warns against binding parties indirectly against their will and emphasizes that transposition of rules should not be confused with systemic integration (Dis. Op. Hmoud, para. 34).

By contrast, Judge Nolte regards the Court’s approach as an “innovative statement” (Sep. Op. Nolte, para. 24) and agrees with the majority’s methodological approach in general. While endorsing the result, he insists that consent remain the anchor for the interpretive means of Art. 31 VCLT. Yet, he diverges from the technical and formalistic understanding of Judges Tomka and Hmoud in favor of a material approach to consent, holding that the consensual element “does not necessarily have to be explicit or formal, it must at least be reflected in a common understanding” (Sep. Op. Nolte, para. 27). Judge Bhandari seems to also endorse this view in his declaration (Decl. Bhandari, para. 8).

Seen against the broader debate on Art. 31(3)(c) VCLT, the majority’s approach is best understood as a deliberate middle course. It rejects the restrictive reading – associated with the World Trade Organization panel in EC – Biotech (paras. 7.68, 7.71) and echoed here by Judges Tomka and Hmoud – under which only rules binding on all parties to the treaty may inform its interpretation. However, the Court stops short of a much more expansive understanding of systemic integration (see e.g. European Court of Human Rights, Demir and Baykara v. Turkey, paras. 69, 78, 86) under which even instruments not binding on the respondent State may inform interpretation.

Unresolved Questions

The standard of a “common understanding” presents an attractive compromise, however, the Court leaves two questions unresolved. First, how silence is to be treated as evidence of that common understanding, and second, how the “relevant rule” is to be distilled from the invoked provisions in the first place.

The Court’s treatment of Saint Lucia’s silence illustrates the first gap as well as the diverging approaches to State silence under Arts. 31 and 32 VCLT. Holding true to the above-described formalist-voluntarist approach, Judge Tomka argues that the majority’s assessment of State silence as affirmative consent disregarded the foundational principle of State consent, as in his view instances in which consent can be inferred from silence are limited (Dis. Op. Tomka, para. 35). Instead, he criticizes that “inaction is repeatedly elevated into affirmative consent” (Dis. Op. Tomka, para. 77). Judge Nolte, by contrast, advances a more nuanced critique that primarily addresses the application of that standard, questioning whether State silence, without any positive indication, can suffice to establish the necessary common understanding. His critique demonstrates his understanding of consent as the guiding criterion in the application of Art. 31 VCLT, yet without elaborating on what would count as an indication of consent. In his opinion, Art. 32 VCLT constitutes the more appropriate method of interpretation to reach the same result, based on the differing degree of State consent both provisions require. As in his view the preparatory work and the circumstances of a treaty’s conclusion only constitute examples of supplementary means of interpretation, Nolte argues for an inclusion of relevant rules not applicable between the parties as an unwritten group of supplementary means (Sep. Op. Nolte, paras. 29, 44). Concerning Art. 32 VCLT, Nolte argues in favour of a theory of “toleration”, which he takes from the ILC’s Fragmentation Report.

This technical argument leads to the same result as the majority’s approach, merely relocating the issue from Art. 31 to Art. 32 VCLT with its diminished normative force. Indeed, silence may amount to acquiescence especially where the circumstances call for a response (Maritime Delimitation in the Indian Ocean (Somalia v. Kenya), ICJ Rep 2021, para. 51). Against this background, the absence of objections by Saint Lucia over decades indeed contribute to the “common understanding”. The opposite view would endow Saint Lucia with the status of a passive bystander in need of protection, rather than a State fully capable of participating in the practice of the treaties it is party to, especially in the relevant proceedings in international courts. Thus, although the Court ultimately reaches a reasonable conclusion regarding State silence, the deeper problem is the absence of any theory of acquiescence. This remains underexplored although the Court had every opportunity to do so.

Turning to the second gap, several judges contested the Court’s derivation of a right to strike from Art. 8(1)(d) and (3) ICESCR, and Art. 22 ICCPR. Noting the joint statement issued in 2019 by the CESCR and the HRC, the Court essentially arrives at the conclusion “that the protection of the right to strike is encompassed in the protection of the freedom of association provided by Convention No. 87” (paras. 97, 98).

Art. 31(3)(c) VCLT requires the Court as interpreter to identify and assess the “relevant” rule of international law. Against this background, both Judges Tomka and Xue criticize the majority’s application of the standard arguing that the Court accorded excessive weight to the Committees’ own interpretations without engaging sufficiently with the treaty texts themselves. Judge Tomka describes this as an act of “double-implication”; Judge Abraham concludes that the references to these interpretations, whose legal weight remained to be established, result in circular reasoning.

While it is certainly true that the inherently vague “relevant” condition affords rather broad interpretive leeway, the Court relied almost exclusively on the treaty bodies’ interpretive output without showing much independent interpretation. The Court does not refer to the interpretive freedoms it took when establishing the benchmark for its engagement with treaty bodies in Ahmadou Sadio Diallo (para. 66) or the subsequent application of this test in its 2021 CERD judgment (para. 101) when citing the CESCR and the HRC. It applies this test only when turning to the ILO supervisory bodies. The case at hand was a prime occasion to engage with the views of treaty bodies in the manner the Court usually does, as only the ICESCR – and not the ICCPR – explicitly refers to the right to strike. Bearing in mind the detailed interpretive approach of the Court in the CERD case, extensively elaborating on the scope of application of CERD and ultimately rejecting the treaty body’s view, the Court’s unexplained refusal to apply its own standard seems unsatisfying and undermines the Court’s arguments under Art. 31(3)(c) VCLT even further. Setting aside the normative content of the rules itself, the criticism convincingly demonstrates that systemic interpretation – apart from missing consensual elements – reaches its limits if the existence or the scope of the “relevant” rule of international law itself cannot be ascertained. Ultimately, a legally controversial rule used for interpretive purposes cannot lend another rule greater normative force than it possesses itself.

Conclusion

The advisory opinion and the judges’ opinions on Art. 31 VCLT are among the most methodologically rich and contested the Court has delivered in recent years. This comes as little surprise as the advisory opinion addresses fundamental questions of treaty interpretation and systemic integration. Concerning the latter, the Court ventured into uncharted territory in choosing an approach resting on vague criteria.

It now falls to the Court to refine these criteria through its practice. Meanwhile, other aspects of systemic integration remain wholly unexplored, such as the relevance of regional human rights treaties for systemic integration (cf. Cleveland) and why the Court chose the controversial “common understanding” approach instead of rekindling the debate on customary human rights law (cf. Cleveland and Hmoud). Bearing in mind that many “constitutions not only incorporate customary international law automatically […] but also grant it a rank superior to that of domestic statutes” (Simma/Alston, p. 86), this opinion seems not as activistly inclined as some might fear. The current opinion seems to favour an approach to systemic integration guided by considerations of effectiveness and toleration, establishing a material understanding of consent. Nevertheless, in failing to spell out these considerations and consequently losing itself in vagueness, the Court unnecessarily leaves itself open to harsh criticism.

*The authors thank Andreas L. Paulus for comments on an early draft.