The Right Result, the Wrong Reasoning?

Does the Freedom of Association and Protection of the Right to Organise Convention, 1948 (Convention No. 87), which never mentions strikes, still protect the right to strike? On May 21, 2026, the International Court of Justice (ICJ) opined by a 10-to-4 majority that it does. The Court reached the right result: excluding the right to strike would render the treaty’s core objective illusory. The advisory opinion strengthens trade unions’ leverage and resolves a longstanding institutional deadlock within the International Labour Organization (ILO). Even so, this piece argues that stronger reasoning was required, consistent with state consent and treaty-interpretation rules. To delve into this further, the blogpost proceeds in two parts. The first analyses the Court’s methodology in the advisory opinion, and the second examines the rules of treaty interpretation and their consent-based foundation.

How the ICJ Found a Right to Strike in Convention No. 87

The Court used Articles 31 and 32 VCLT as its interpretive framework. Although Convention No. 87 predates the VCLT, Articles 31 and 32 still apply as customary international law. The Court acknowledged the Convention’s silence on the right to strike; nonetheless, the Convention’s broader scheme and purpose supported its inclusion (¶68).

First, applying Article 31(1) of the VCLT, the Court acknowledged that Convention No. 87 does not expressly mention the right to strike; however, this silence is inconclusive (¶67–68). Article 2 of Convention No. 87 lays down the right to form or join unions, and Article 10 defines ‘organisations’ as a means of “furthering or defending the interests of the workers” (¶69). Alongside this, Article 3(1) allows them to “organise their administration and activities and to formulate their programmes.” While these terms are not defined, colloquially, ‘activities’ are actions taken to reach a goal. Reading these provisions together, the Court concluded that the strike falls within the broader ambit of ‘activities’ and that this interpretation is reinforced by the Convention’s objective of guaranteeing freedom of association and improving labour conditions and rights (¶72).

Second, Article 31(3)(a) VCLT requires courts to consider ‘subsequent agreements’ between the parties, but no such agreement between the state parties was found in this case. Similarly, the states’ conflicting views regarding the interpretation indicated a lack of consensus, and thus no ‘subsequent practice’ among the states under Article 31(3)(b) could be established (¶87). Therefore, the Court relied on Article 31(3)(c), which guides treaty interpretation in the context of relevant rules of international law. Consequently, it turned to Article 8 of the International Covenant on Economic, Social and Cultural Rights (ICESCR), which protects the right to strike. Additionally, the Court considered the interpretation of Article 22 of the International Covenant on Civil and Political Rights (ICCPR) in human rights practice, which treats the right to strike as an important aspect of freedom of association. The Court also considered the 2015 Government Group statement, in which 28 representatives recognised that the right to strike forms part of freedom of association.

Finally, applying Article 32 of the VCLT, the ICJ considered supplementary means of interpretation. First, the travaux préparatoires (preparatory works) proved inconclusive, as they neither supported nor excluded the right to strike, leaving the issue ambiguous. Second, the Court found that the subsequent practice did not establish the agreement of the parties under Article 31(3)(b) as an ‘authentic means of interpretation’. It nevertheless considered the same practice under Article 32 as a supplementary means of interpretation to confirm, rather than establish, its interpretation that Convention No. 87 protects the right to strike (¶112–115). Third, the Court ascribed great weight to supervisory bodies, namely the Committee of Experts on the Application of Conventions and Recommendations (CEACR) and the Committee on Freedom of Association (CFA), as well as to regional instruments from different regions, such as the African Charter on Human and Peoples’ Rights (¶41-50).

Interpretation and State Consent

The dissenters identified a methodological flaw in the majority’s reasoning. Judge Tomka noted that the majority had overlooked the context that prompted the Convention’s drafting (dissenting op. ¶15). Judge Abraham advanced a similar critique: the question was not whether the right to strike existed, but whether states had consented to include it in Convention No. 87 (dissenting op. ¶9). By bypassing this question, the majority thus blurred the distinction between interpretation and supplementing a treaty.

Treaty interpretation under the VCLT is not about finding a hidden meaning but about attributing the meaning. Scholars like Christian Djeffal describe this as a “2.5 step hierarchy” according primary weight to Article 31 indicators while reserving the travaux préparatoires under Article 32 for a supplementary step (the “0.5 step”). The VCLT rejects rigid textualism of meaning, and this semantic flexibility enabled the ICJ majority to expand the ordinary meaning of ‘activities’ under Article 3(1) of Convention No. 87. However, the International Law Commission (ILC), in its 1966 Draft Articles on the Law of Treaties and accompanying commentaries, acknowledged that interpretative discretion cannot be unlimited. The text must be interpreted based on the circumstances of the case, and interpretation cannot serve as a means of revising a treaty or adding a meaning that was originally absent (p. 221, ¶11). To illustrate the risk: imagine two states that jointly draft a treaty clause requiring ‘reasonable cooperation’ on security without ever discussing mandatory intelligence sharing. Later, a court interprets the provision’s textual silence to add new legal duties, such as requiring oversight committees to meet every two months. This is precisely the interpretative overreach that characterises the ICJ opinion.

The foundational anchor of this interpretative approach lies in Article 31(1) of the VCLT, which states:

“A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.”

This sequence matters. Ordinary meaning constrains the interpretative process. Context under Article 31(2), read with the treaty’s object and purpose, guides the interpreter in determining the meaning. Article 31(3) then further admits three considerations: subsequent agreements, subsequent practice, and relevant rules of international law. Article 32 reserves travaux préparatoires and the circumstances of the Convention’s conclusion as supplementary means of interpretation. Courts may use them only to confirm an Article 31 interpretation or to resolve genuine ambiguity. Crucially, the ICJ emphasises that these elements operate as a “single combined operation” (ICJ Climate Change Advisory Opinion, ¶177).

The present case exposes this tension. Article 31(3)(a),(b) focus on state consent. Yet the majority relied on systemic integration under Article 31(3)(c), focusing on external norms (¶75, ¶86–89). It reinforced this conclusion with supplementary means under Article 32, reading the term ‘including’ as non-exhaustive (¶102). Notably, the majority itself conceded that the Convention does not answer the question of the status of the right to strike (¶110). The Court therefore needed to explain the scope of interpretation; herein, the Court simply imported external norms to fill textual silence. The VCLT permits evolutive interpretation, yet the principal indicators of the parties’ agreement are absent, namely subsequent agreement and practice. Article 31(3)(a) and (b) further need to be read in the context of the Convention’s object and purpose and state consent. By contrast, systemic integration under Article 31(3)(c) relies on external rules to be ‘applicable in the relations between the parties’. The majority herein concluded common understanding ‘may be presumed’ from high ratification rates of the ICESCR and ICCPR (¶90, ¶92). Yet four state parties to Convention No. 87 are not parties to both instruments, while certain other states had entered reservations regarding the ICESCR (¶94–95). Thereby greater scrutiny was required for application of Article 31(3)(c) VCLT. Additionally, the Court also treated parties’ silence as implied consent despite divergent domestic labour laws (Judge Tomka, dissenting op. ¶79).

Notably, this problem is amplified by the centrality of state consent to treaties. Treaties bind successive governments as an enduring legal obligation; thereby, courts must exercise restraint rather than expanding treaty commitments beyond what was originally accepted. The ILC Commentary emphasised that the text itself is presumed to be the ‘authentic expression of the intentions of the parties’, establishing that treaty interpretation must focus on elucidating the text rather than searching for unexpressed intent (Draft Articles on the Law of the Treaties with commentaries, 1966, page 223, ¶18). President Iwasawa similarly stated: “Authentic interpretation means the interpretation agreed upon by all parties to the treaty” (separate op. ¶12). Subsequent agreements and subsequent practice therefore carry particular weight because they reflect the parties’ shared understanding regarding the treaty’s meaning. Their interpretive authority derives from the parties’ continuing consent.

Beyond Advisory Opinion

Based on this discussion, first, neither the Convention nor the VCLT treats textual silence as a bar to interpretation. However, Article 31 VCLT is designed primarily to elucidate the authentic meaning of the agreed text, not to serve as an automatic vehicle for evolutive adaptation sans state consent. An evolutionary interpretation cannot amount to treaty revision when a text is silent. Unlike textual ambiguity, where an interpreter chooses between competing meanings already contained in the text, silence concerns the absence of an express rule. The VCLT framework proceeds from the treaty text, which is presumed to be the authentic expression of the parties’ intentions. Consequently, silence should be approached with caution and relied upon only where necessary to give effect to the treaty within the interpretive framework of the VCLT. In sum, evolutive interpretation cannot create wholly new obligations. In the present case, the majority opinion relied on the ICCPR and the ICESCR as evidence of a “common understanding between parties”, which is erroneous. Judge Tomka also observed that several states like Bangladesh, Ireland, and Switzerland have expressly objected to the right to strike in the Convention (dissenting op. ¶35-39). Nor do the ICCPR and the ICESCR reflect universal consensus, given relevant reservations. Furthermore, tripartite constituents (government, workers, and employers) negotiated Convention No. 87, making its drafting process significantly different from human rights treaty-making. In this tripartite setting, identifying state consensus is difficult. Thus, the Court had greater reason to exercise caution before importing external human rights treaties to fill textual silence.

Second, as the Court couldn’t find any common understanding under Article 31(3)(a) and (b), the Court relied primarily on external legal rules under Article 31(3)(c) to support its interpretation. Article 31 VCLT must be read as a single integrated framework as aforesaid. Despite this, prioritising external rules over state practice sidelines the contracting states’ consent. Article 31(3)(a) and (b) concern authentic means of interpretation, as they reflect the parties’ original consent.  President Iwasawa also questioned whether an interpretation based on Article 31(3)(c) has the same legal force as an “authentic” interpretation as one reached under (a) and (b) (separate op. ¶14). The ILC Draft Conclusions on Subsequent Agreements and Subsequent Practice observes that it can be difficult to establish the consent of all state parties to an expert treaty body’s interpretation (page 84, ¶12). Judge Hmoud too states the majority did not justify treating a 28-member group statement as a representation of a significant majority. State silence does not amount to endorsement, especially when state parties do not expect their passivity to bind them to a specific interpretation (Judge Hmoud, dissenting op. ¶25).

Third, the majority opinion relied on good faith, sidelining drafters’ intention. Article 31 indeed requires a treaty to be interpreted in good faith; however, the principle of good faith does not authorise courts to supplement treaties with desirable obligations. The majority instead adopted a broader understanding of good faith in light of Convention No. 87’s object and purpose.

Conclusion

Ultimately, the advisory opinion resolves a genuine institutional crisis and protects workers worldwide. However, the outcome must not be conflated with the Court’s reasoning. The right to strike should ideally be derived from the Convention’s text, context and object and purpose. The Court erred by using Article 31(3) VCLT, which focuses on state consent, and by misapplying Article 32 as a supplementary means, it bypassed primary sources. Persistent objections by states make reliance on external instruments questionable.

In the end, the Court’s eagerness to deliver a progressive outcome led it to forego methodological rigor, leaving a doctrinal hole in its interpretative framework, which might be followed in the future. If the provision was unclear, the Court should have refined those interpretive boundaries. Specifically, the Court must clarify the methodological limits of Article 31(3)(c) where there are state reservations or persistent objections. Alongside this, the ICJ must articulate a clearer distinction between resolving textual ambiguity and filling textual silence. The opinion left two critical aspects unanswered: first, how external treaties can denote a common understanding where the text is silent, and second, how and in what circumstances silence in a treaty should be interpreted. While treaties must accommodate legal developments with dynamic socio-economic reality, evolutive interpretation cannot be used to bypass sovereign consent.

None of this suggests that silence should foreclose interpretation altogether, since silence is an inherent feature of negotiated multilateral treaties. Rather, the argument here is that where courts seek to attribute legal consequences to textual silence, the justificatory burden correspondingly increases. As the interpretive basis moves further away from the express text and direct manifestations of parties’ agreement, stronger reasons must be offered to demonstrate that the resulting interpretation remains anchored in the consensual foundations of international law.