Sources — Treaties (the VCLT 1969 regime)

1. FRAME — the consent paradox

Article 38(1)(a) ICJ Statute lists "international conventions, whether general or particular, establishing rules expressly recognised by the contesting states" as the first formal source. The wording embeds the consensualist premise: a state is bound only by what it has accepted. The VCLT — adopted at the Vienna Conference 1968–69, in force 1980 — codifies that premise; the ICJ treats most provisions as declaratory of customary law, binding non-parties for the customary content.

Two propositions sit in tension. Treaties rest on free consent: Article 26 (pacta sunt servanda); Article 34 ("a treaty does not create either obligations or rights for a third State without its consent"). Yet consent is not unlimited: it is policed by reservations (Articles 19–21), interpretation (Articles 31–32), invalidity (Articles 51–52, 60), and peremptory norms (Articles 53, 64). The exam question lives in that gap. Examiners reward IRAC structure, specific authority, debate engagement, and a sustained thesis; they penalise long historical introductions and footnotes.

2. EXPLAIN — the VCLT architecture

Five blocks carry the doctrinal load.

• Definition and consent (Articles 2, 11–17). Article 2(1)(a): a treaty is "an international agreement concluded between States in written form and governed by international law … whatever its particular designation". Form is irrelevant; legal effect is the test. Article 11 lists the means of consent — signature, ratification, acceptance, approval, accession.
• Reservations (Articles 2(1)(d), 19–23). Article 2(1)(d): "a unilateral statement, however phrased or named … where it purports to exclude or to modify the legal effect of certain provisions". Article 19 fixes the outer limit — impermissible if prohibited, outside a specified-reservations clause, or "incompatible with the object and purpose of the treaty" (Article 19(c)). Articles 20–21 operate the bilateral acceptance/objection machinery.
• Performance and interpretation (Articles 26, 31–32). Article 31(1): "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". Article 32 admits supplementary means (travaux) only to confirm Article 31 or to resolve ambiguity, obscurity, or manifestly absurd results.
• Invalidity (Articles 48, 51–52, 53, 60, 64). Error; coercion of a representative; coercion of a state by force; conflict with jus cogens; material breach; supervening peremptory norm. Article 52 voids any treaty "procured by the threat or use of force … embodied in the Charter" — orthodoxy reads "force" as armed force.
• Third parties and peremptory norms (Articles 34–37, 53, 64). Article 34 codifies pacta tertiis nec nocent nec prosunt. Article 53 voids treaties conflicting with a norm "accepted and recognised by the international community of States as a whole" from which no derogation is permitted; Article 64 voids existing treaties on emergence of a new such norm.

Treaty/custom (Topic 02) runs through the regime: the VCLT itself has crystallised into custom (applied to non-parties in Gabčíkovo-Nagymaros 1997, Pulp Mills 2010); treaty rules may also generate custom (North Sea Continental Shelf 1969).

3. The starred cases & instruments

4. CRITIQUE — pressure points

1. Bilateral regime, multilateral object. Article 20 generates pairwise relationships — fine for synallagmatic obligations, poor fit for an erga omnes partes obligation like CEDAW Article 2 or ICCPR non-derogation. The lecturer's "Holy Book" reservation to CEDAW Article 2 almost certainly impairs the raison d'être — but who decides, with what consequence?
2. Who polices Article 19(c)? Belilos (1988) and HRC GC 24 (1994) assert inherent competence in supervisory bodies. The ILC Guide 2011 splits the difference: para 4.5.1 makes incompatible reservations null and void; para 4.5.3 presumes severance. DRC v Rwanda (2006) keeps the ICJ narrow — Rwanda's reservation to Article IX did not impair object and purpose because it touched only dispute settlement.
3. Article 31 in bad faith. The Bybee Torture Memo (1 August 2002) re-defined "severe pain" in Article 1 of the Torture Convention 1984 as pain "akin to that which accompanies … organ failure, impairment of bodily function, or even death". Against Article 31, textbook bad faith; travaux (Article 32) confirm a much lower threshold. The standard worked example for the limits of textualism when an executive instrumentalises Article 31.
4. Article 52 and unequal treaties. Orthodoxy reads "force" as armed force. Post-colonial states argued for a wider reading capturing the 19th-century treaties imposed on China (Nanking 1842, Tianjin 1858) and modern economic conditionality, sanctions, aid withdrawal, tariff threats. Vienna rejected the wider reading but adopted the Declaration on the Prohibition of Military, Political or Economic Coercion as a rider. Chimni reads narrow Article 52 as an imperial framing of consent.
5. Jus cogens — safety valve or politics? Article 53 voids treaties conflicting with peremptory norms; Conclusion 23 (2022) supplies a non-exhaustive list. The catalogue is contested. D'Amato (1990) calls "no derogation permitted" circular. Koskenniemi argues the list reflects post-1945 liberal-internationalist preferences. Charlesworth notes the absence of gender-based discrimination — a feminist critique of which interests are coded "fundamental".

5. EVALUATE — taking a thesis

Three theses; say which you adopt and why.

Thesis A — codifying achievement. Following Aust and Crawford, the VCLT is progressive development codified into custom. Its rules survive because they balance consent against pragmatic recognition that treaties operate over long horizons; the Article 19–21 compromise is the price of universal participation, and Belilos and the HRC show the regime supple enough to accommodate adaptation.

Thesis B — structural mismatch. Designed in 1969 for bilateral, synallagmatic agreements; ill-fitted to the multilateral human rights, environmental and disarmament treaties that have dominated since. Bilateral acceptance/objection, the narrow Article 19(c) reading in DRC v Rwanda, and the absence of centralised admissibility-policing point to a regime protecting sovereignty at the cost of community norms. Belilos, GC 24 and ILC Guide para 4.5 are workarounds, not solutions.

Thesis C — consent critique. Following Chimni's TWAIL reading, the VCLT codifies a specifically imperial consent. Narrow Article 52 immunises 19th-century unequal treaties; jus cogens, as Koskenniemi and Charlesworth show, encodes post-1945 liberal preferences while excluding gender, economic and ecological hierarchies. Not neutral machinery; a particular settlement of power.

The defensible position holds A's codifying achievement together with B's diagnosis of structural strain, with C as a brake on triumphalism. The VCLT is the indispensable starting point — but the modern multilateral treaty has outgrown its original frame, and reform happens piecemeal through human rights bodies, ILC Guides and judicial interpretation, not amendment.

6. Academic voices

The Vienna Convention is a workable, broadly accepted, largely customary codification. Its rules on reservations and interpretation have proved more flexible in practice than their text suggests, depending on whether states and tribunals apply them in good faith. (paraphrased)

— Anthony Aust, Modern Treaty Law and Practice (CUP, 3rd edn 2013); cf. Catherine Brölmann, "Law-Making Treaties" (2005); James Crawford, Brownlie's Principles (9th edn 2019), ch 16.

"It's a bird, it's a plane, it's a jus cogens norm!" — the rhetorical inflation of peremptory norms has run far ahead of any agreed method for identifying them. The Article 53 criterion — a norm "from which no derogation is permitted" — is circular: how do we know derogation is impermissible except by inspecting the very practice we are trying to evaluate? (paraphrased)

— Anthony D'Amato, "It's a Bird, It's a Plane, It's Jus Cogens!" (1990) 6 Connecticut Journal of International Law 1.

Hierarchy in international law is never simply a matter of identifying pre-existing peremptory norms; the catalogue of jus cogens is a political artefact reflecting post-1945 liberal-internationalist priorities. To call a norm peremptory is to take a side, not to describe a fact. (paraphrased)

— Martti Koskenniemi, "Hierarchy in International Law: A Sketch" (1997) 8 EJIL 566; cf. Bruno Simma & Andreas Paulus, "The 'International Community'" (1998) 9 EJIL 266.

The VCLT codified a Eurocentric law of treaties at the moment newly independent states were arguing for a wider concept of coercion to capture the unequal treaties imposed on China and the Ottoman Empire and the economic conditionality of post-colonial relations. Article 52's narrow "force" is not technical neutrality; it is the imperial settlement of the consent question. The canonical jus cogens list — silent on gender — confirms whose interests the doctrine has been built to protect. (paraphrased)

— B.S. Chimni, International Law and World Order (CUP, 2nd edn 2017); cf. Hilary Charlesworth & Christine Chinkin, The Boundaries of International Law: A Feminist Analysis (Manchester UP, 2000).

7. Common pitfalls

8. Exam answer skeleton

1. FRAME (Issue). Identify the proposition; pin the consent paradox (VCLT codifies consent in Articles 26, 34 and polices it in 19, 31, 52, 53); state thesis in one sentence.
2. EXPLAIN (Rule). VCLT 1969, mostly customary. Anchor: Articles 2(1)(a), 11–17, 19–21, 26, 31–32, 34, 52, 53, 64. Reservations to the Genocide Convention AO 1951 as the hinge.
3. CRITIQUE (Application). Bilateral regime vs human rights treaties; who polices Article 19(c) (Belilos 1988, GC 24, ILC Guide 2011, DRC v Rwanda 2006); bad-faith interpretation (Bybee 2002); Article 52 and unequal treaties (Chimni); jus cogens as politics (D'Amato, Koskenniemi, Charlesworth).
4. EVALUATE. Thesis A (codifying achievement) / B (structural mismatch) / C (consent critique). A+B with C as brake.
5. CONCLUDE. Take a position: the VCLT is the indispensable starting point; the modern multilateral treaty has outgrown its frame; reform happens piecemeal, not by amendment.

9. Checklist

• IRAC structure at the top; thesis in one sentence.
• Anchored every claim in a VCLT article (2, 11, 19, 20, 21, 26, 31, 32, 34, 52, 53, 64).
• Reservations to the Genocide Convention AO 1951 for the compatibility test.
• Belilos v Switzerland (1988) for nomenclature-irrelevance and severance.
• DRC v Rwanda (2006) paras 64–70 for the narrow Article 19(c) reading and the substantive/procedural distinction.
• Bybee Memo (2002) as the worked bad-faith Article 31 example.
• Unequal-treaties debate with named authority (Chimni 2017).
• ILC 2022 jus cogens annex and at least one sceptic (D'Amato, Koskenniemi or Charlesworth).
• Took a position. No undefended description; no long historical introductions; no footnotes.