Can the ILO Interpret Differently?

During the drafting of the Vienna Convention on the Law of Treaties, 1969 (‘VCLT’ or ‘the Convention’), Sir Humphrey Waldock famously remarked that treaty interpretation is “to some extent an art, not an exact science”. Almost six decades later, the operationalisation of this principle has, once again, come before the International Court of Justice (‘ICJ’) in its advisory opinion on the ‘Right to Strike under ILO Convention No. 87’. Following the filing of fifteen written comments from States and organizations and public hearings, the ICJ, on May 21, 2026, declared that the ILO Convention on Freedom of Association and Protection of the Right to Organise, 1948 (No. 87) protects the right to strike of workers and their organizations. Among the issues before the Court was the technical question of whether the ILO’s instruments constitute lex specialis under Article 5 of the VCLT, thereby modifying the Convention’s traditional interpretative framework. This piece critiques the ICJ’s treatment of this question and confines its analysis to this technical issue of treaty interpretation.

Article 5 of the VCLT: Nature and Threshold

Article 5 of the VCLT provides that “the present Convention applies to any treaty which is the constituent instrument of an international organization and to any treaty adopted within an international organization without prejudice to any relevant rules of the organization” [emphasis added]. Anderson and Wood note that “without prejudice” is the key term in Article 5. In Fisheries Jurisdiction, the ICJ interpreted this term, used in the 1961 Exchange of Notes Agreement, as a savings clause that qualifies the operative provision while preserving its effect (paragraph 37). Accordingly, where an international organization contains separate and distinct rules on a particular subject of treaty law, those rules take priority over the Convention.

The rationale behind the inclusion of this provision was to ensure that the VCLT’s rules do not encroach upon the internal legal regimes of international organizations (page 161). This was confirmed by Vienna Conference participant and ICJ Judge de Castro in his separate opinion in the ICAO Council Appeal (India v. Pakistan) case, who observed that the discussions at Vienna had emphasised the “need for utmost respect for the peculiarities of international organizations” and observed that States at the Vienna Conference did not wish to weaken the growing achievements and effectiveness of such organizations (page 129). Similarly, Mark Villiger notes, “one aim of the Conference was not to impair the rights enjoyed by international organisations”.

At the same time, it was clarified during the drafting process that Article 5 was not intended to operate as a general reservation clause in favour of international organizations (pages 56-57). As a result, Article 5 does not permit a complete exemption from the framework of the VCLT, but instead operates on a provision-by-provision basis. It allows for the displacement of specific VCLT provisions where a “relevant rule of the organization” exists. Such rules may include written rules, unwritten customary rules, or even established institutional practices, provided that such practices are mandatory and accompanied by opinio juris (page 147). This understanding of “rules of the organization” is also reflected in Article 2(1)(j) of the VCLT, 1986.

The Contention Before the ICJ

The provision whose displacement by Article 5 was at issue was Article 32. Through Articles 31 and 32, the VCLT has been understood to establish a hierarchy between the ordinary meaning of the text and its travaux préparatoires, with the latter functioning only as a supplementary means of interpretation. However, in the course of the written and oral pleadings before the ICJ, the International Labour Office (‘Office’) and the International Organisation of Employers (‘IOE’), with the support of Switzerland, and Business Africa, had argued that recourse to preparatory work within the ILO framework is not subject to such hierarchy. According to this view, the ILO’s particular reliance on preparatory work constitutes a “relevant rule of the organization” such that the preparatory work of ILO conventions would receive greater interpretative weight than that assigned to it as a supplementary means under Article 32. On the other hand, the International Trade Union Confederation (‘ITUC’), on behalf of employees, denounced this “backdoor” entry of Article 32 extra-textual materials in the absence of any ambiguity arising from the textual interpretation of ILO Convention No. 87.

The ICJ, in its Advisory Opinion, only briefly addressed this question and was not persuaded that the interpretative approach developed through the practice of ILO bodies constituted “a relevant rule of the organization” capable of affecting the application of the customary international law rules of treaty interpretation (Paragraphs 63-65). However, Justice Bhandari’s Separate Opinion engages with the issue in greater detail. He concurs that the principle of lex specialis does not apply, reasoning that Article 5 of the VCLT would apply to Convention No. 87 only insofar as it reflected customary international law. Further, he relies on the ILO Committee of Experts on the Application of Conventions and Recommendations (“CEACR”) reference to Article 32 as a supplementary means of interpretation (paragraph 10). He notes that the Office’s opinion was the sole authority supporting the existence of a lex specialis norm. However, he argues that a norm of such significance cannot be sustained in the absence of sufficient support from either ILO Member States or consistent institutional practice (paragraph 11).

This piece critiques the approach adopted by the ICJ and examines whether the ILO’s recourse to preparatory work constitutes a “relevant rule” within the meaning of Article 5. It then addresses the extent to which the application of Article 5 would alter the traditional interpretative relationship between Articles 31 and 32.

ILO’s Recourse to Preparatory Work as a “Relevant Rule”

In order to invoke Article 5, the practice in question must amount to a “relevant rule of the organization”. As noted above, this concept encompasses any rule that has been followed over a sufficiently long period and is accompanied by the requisite opinio juris.

In 1952, the ILO Director-General issued Instruction No. 45 concerning the procedure for responding to requests for interpretations of Conventions and Recommendations. The Instruction required the Office to consult the preparatory work of the relevant instrument when preparing the memorandum, as well as to cite it explicitly in the response. Similar instructions were reiterated in the 1968 Instruction No. 337 and a 1987 circular.

Furthermore, in 1993, the Office prepared a document on interpretation for the Governing Body, in which it underscored the importance of the intention of the parties in interpretive methodology. It emphasised that Article 5 could be invoked to highlight the significance of preparatory work in ILO practice (paragraph 46). Similarly, in 2009, the International Labour Standards Department, in consultation with the Office of the Legal Adviser, prepared a ‘non-paper’ on the interpretation of ILO Conventions. This document again referred to Article 5 and explained how the interpretative practices of the ILO’s supervisory bodies attach significant weight to preparatory work (paragraphs 22-26).

Since this practice has continued even after the adoption and entry into force of the VCLT, former ILO Legal Adviser Anne Trebilcock observes that “little or no impact of the VCLT can be seen in ILO policy on (unofficial) interpretations issued by the Secretariat”. This is in line with ILO being credited as the key driving force behind the adoption and formulation of Article 5. At the Vienna Conference, Jenks characterised the ILO’s recourse to preparatory work as a practice not aligned with the VCLT framework (Pages 36-37).

In light of these institutional directions issued by the Director-General, together with the consistent conduct of ILO actors and bodies tasked with applying and clarifying ILO Conventions, Justice Bhandari’s conclusion that institutional practice does not support according primacy to travaux préparatoires is difficult to sustain. In fact, his reliance on the CEACR’s reference to Article 32 is misplaced (paragraph 11). At the 100th Session of the International Labour Conference, the CEACR observed that “the Committee has always paid due regard to the textual meaning of the words in light of….Article 31 of the Vienna Convention…..In addition and in accordance with Articles 5 and 32, the Committee takes into account the Organization’s practice of examining the preparatory work leading to the adoption of the Convention…” (paragraph 12) [emphasis added]. Thus, the CEACR did not treat Article 32 as merely supplementary; rather, it considered preparatory work alongside Article 31 pursuant to Article 5.

The next part considers how recognising the ILO’s interpretative practice as lex specialis under Article 5 would have affected the Court’s interpretative methodology.

The Role of Article 5 in Modifying the Traditional Rules of Interpretation

As noted above, Article 32 permits recourse to supplementary means of interpretation, but limits such recourse to two situations, namely, to confirm an interpretation arrived at under Article 31, or to determine meaning where the Article 31 interpretation leaves the text ambiguous or obscure, or leads to a manifestly absurd or unreasonable result. The arguments advanced by the Office and the IOE do not challenge the primacy of Article 31. Interpretation of ILO Convention No. 87 must still begin with the ordinary meaning of the text. What is contested, however, is the role of Article 32. Specifically, it is argued that within the ILO framework, Article 32 must be accorded a co-primary status and considered alongside Article 31, rather than being confined to a subsidiary role triggered only upon ambiguity or absurdity. This shift is methodological in nature as Article 5 operates to remove the threshold requirement that the textual interpretation must first yield ambiguity or absurdity before recourse to preparatory work is permitted.

It then becomes necessary to consider whether this shift has any practical significance. The author argues that the gap between Articles 31 and 32 is, in any event, already narrow. Since Article 32 permits recourse to preparatory work to “confirm” meaning, Gardiner rightly observes that such confirmation inherently carries the possibility of non-confirmation. In other words, an examination of preparatory work may lead the interpreter to revisit the textual analysis and adopt a different, yet still permissible, interpretation. Similarly, recourse to preparatory work may reveal ambiguities that were not initially apparent, thereby transforming its role from confirmatory to determinative. As Dörr, and Schmalenbach note, it is in practice difficult to conceive of situations in which preparatory work is entirely excluded from interpretation.

This is most evident in the Advisory Opinion itself. The Court observed that the textual interpretation of Convention No. 87 under Article 31 did not produce an interpretation that was ambiguous or obscure, nor one that led to a manifestly absurd or unreasonable result. Nevertheless, it proceeded to rely on supplementary means of interpretation under Article 32 in order to ascertain whether they confirmed the meaning derived through Article 31 (paragraph 101). This approach is consistent with the Court’s reasoning in the CERD (Georgia v. Russia) case, where the ICJ observed that, although it was unnecessary to resort to supplementary means in order to determine the meaning of Article 22 of CERD, the extensive arguments advanced by the parties concerning the travaux préparatoires warranted their examination (Paragraph 142).

Consequently, since the distinction between Articles 31 and 32 is already limited in practice, and given that the ICJ itself ultimately resorted to Article 32, the Court’s failure to meaningfully engage with Article 5 appears methodologically unpersuasive.  It is true that recognition of the ILO’s interpretative practice as lex specialis may not have produced a different substantive outcome in this case; however, it would have altered the methodology by which the Court approached preparatory work. The significance of Article 5 therefore lies in establishing the legal basis for according preparatory work a status beyond the supplementary role contemplated by Article 32, while also clarifying the scope of interpretative autonomy available to institutional regimes such as the ILO.

Conclusion

This piece maps the ongoing question of whether the reliance by the Office and the IOE on Article 5 of the VCLT, to accord lex specialis status to preparatory work over its traditionally supplementary role, is justified. An analysis of the arguments advanced by all parties, the interpretative practice of ILO organs and the nature and scope of Article 5 suggests that such reliance is indeed justified. At the same time, recognising this practice as lex specialis would not necessarily have produced a different outcome in the Advisory Opinion. The significance of the issue therefore lies less in the substantive outcome of the Advisory Opinion and more in the methodological coherence of the Court’s approach to treaty interpretation within institutional legal regimes such as the ILO.