Module 1: Sources & Core Doctrines

Article 38 of the ICJ Statute is the traditional catalog of international law sources: treaties, custom, and general principles of law recognized across major legal systems, plus, as subsidiary means, judicial decisions and the writings of highly qualified publicists (scholars) — evidence of the law, not independent binding sources in themselves.

Example. Peteropolis and John-land sign a formal boundary treaty — binding by treaty. Separately, states have long treated a particular navigation rule as legally required even absent any treaty — binding by custom, if the elements are met. An international tribunal, lacking either, might apply a widely shared domestic-law principle like good faith — a general principle filling the gap.

Article 38’s text is old enough to predate the United Nations itself. It first appeared, in nearly identical wording, in the 1920 Statute of the Permanent Court of International Justice — the League of Nations’ own world court — and was carried forward essentially unchanged into the ICJ Statute when the United Nations replaced the League after World War II. That continuity is itself a real, useful data point: the basic catalog of where international law comes from has remained remarkably stable for over a century, even as the institutions applying it, and the substantive rules found within each source, have changed enormously. When courts and tribunals today cite “Article 38” as the traditional starting point for identifying a rule of international law, they are relying on language largely unchanged since 1920.

Customary international law requires two elements: state practice that is genuinely widespread and consistent, and opinio juris — a sense that the practice is legally required, not merely convenient or courteous. The persistent objector rule can let a state that clearly and consistently objected from a norm’s early formation escape being bound by it (though this generally doesn’t apply to jus cogens norms).

Example. Dozens of states, over decades, refrain from taxing emergency famine relief shipments, repeatedly describing this publicly as legally required — likely binding custom, both elements satisfied. Now compare: states have long waved ships through a busy strait purely as a courtesy, with no suggestion it’s legally required — mere comity, since opinio juris is missing, even though the practice itself is widespread.

The North Sea Continental Shelf Cases (Federal Republic of Germany v. Denmark; Federal Republic of Germany v. Netherlands), 1969 I.C.J. 3, is the foundational real case establishing this two-element test in practice. Denmark and the Netherlands argued that the “equidistance principle” for delimiting continental shelf boundaries — found in the 1958 Geneva Convention on the Continental Shelf, which Germany had never ratified — nonetheless bound Germany as customary international law. The International Court of Justice disagreed, holding the rule hadn’t crystallized into binding custom: state practice applying equidistance wasn’t yet sufficiently widespread and uniform, and critically, the states that did apply it did so largely because the treaty required them to, not out of any independent sense of legal obligation — meaning opinio juris was genuinely missing. North Sea is the reason this section insists both elements are independently required: widespread practice alone, without evidence states believed themselves legally bound rather than merely following a treaty or acting out of convenience, doesn’t create binding custom.

The Vienna Convention on the Law of Treaties (VCLT) — the “treaty on treaties” — governs how binding agreements between states actually work. Pacta sunt servanda (“agreements must be kept”) is the foundational principle that treaties in force bind the parties, who must perform them in good faith.

Example. Peteropolis and John-land sign and ratify a fishing rights treaty. Once in force, both states are bound to honor it in good faith under pacta sunt servanda — Peteropolis can’t simply ignore inconvenient provisions later without following the treaty’s own rules (or the VCLT’s) for suspension, amendment, or termination.

The Gabčíkovo-Nagymaros Project Case (Hungary v. Slovakia), 1997 I.C.J. 7, shows pacta sunt servanda applied to real, high-stakes facts. Hungary and Czechoslovakia had signed a 1977 treaty to jointly build a major dam and hydroelectric system on the Danube. Hungary later suspended and then purported to terminate its obligations, citing environmental concerns that arose after the treaty was signed, while Slovakia (Czechoslovakia’s successor) proceeded with an alternative version of the project on its own side of the border. The ICJ held Hungary’s unilateral suspension and termination were unlawful: a state cannot simply invoke changed circumstances or its own later environmental judgment to excuse itself from a validly concluded treaty still capable of performance — it must instead use the treaty’s own dispute-resolution mechanisms or the VCLT’s own narrow, specifically defined grounds for termination. Gabčíkovo-Nagymaros is a genuinely concrete illustration of how demanding pacta sunt servanda actually is in practice, even against a state’s sincerely held later objections.

A reservation is invalid if prohibited by the treaty, outside a specified permitted list, or incompatible with the treaty’s object and purpose. Treaties are interpreted in good faith, per their ordinary meaning in context and the treaty’s purpose (VCLT Article 31). A material breach can give grounds for suspension or termination (Article 60).

Example. A state ratifies a human rights treaty but reserves entirely out of its central non-discrimination provision — likely invalid, since it guts a core commitment. Now compare: a state instead reserves out of one narrow, peripheral reporting deadline — likely valid, since it doesn’t touch the treaty’s essential purpose.

The object-and-purpose test itself has a real, specific origin: the ICJ’s Reservations to the Genocide Convention advisory opinion, 1951 I.C.J. 15. Several states had ratified the 1948 Genocide Convention with reservations, and other parties objected, raising a genuinely open question at the time: does a single objection from one party block a reserving state from becoming a party at all, following the older, strict rule requiring unanimous consent to any reservation? The Court broke from that older rule, holding a reserving state could still be considered a party, vis-à-vis states that didn’t object, so long as the reservation was compatible with the convention’s object and purpose — a flexible, case-by-case standard rather than a rigid unanimity requirement. That 1951 opinion is the direct doctrinal ancestor of the VCLT’s own object-and-purpose language, drafted nearly two decades later specifically to codify the more flexible approach the Genocide Convention opinion had pioneered.

Sovereign equality (UN Charter Article 2(1)) means every state holds equal legal status, regardless of its actual size, power, or wealth — a legal principle about formal standing, not a factual claim about equal real-world influence. Territorial sovereignty gives a state exclusive authority to govern within its own territory, a foundational premise of the whole system.

Example. A small island nation and a major world power are formally equal parties before the International Court of Justice, each with one vote and equal legal standing to bring or defend a case — sovereign equality in action, even though the two states are wildly different in actual size and power.

The Island of Palmas Case (Netherlands v. United States), 2 R.I.A.A. 829 (1928), remains the foundational arbitration on what territorial sovereignty actually requires. Both the Netherlands and the United States claimed the small island of Palmas (Miangas) — the U.S. based its claim on Spain’s earlier discovery and cession of the Philippines (which it argued included the island), the Netherlands on its own long-standing, actual exercise of authority there through the Dutch East India Company and later colonial administration. Sole arbitrator Max Huber ruled for the Netherlands, holding that a mere historical claim of discovery, without continuous and actual display of state authority afterward, could not establish or maintain sovereignty against a rival state that had actually and peacefully exercised authority over the territory for a sustained period. Palmas is the reason territorial sovereignty is understood as a matter of genuine, ongoing effective control, not simply a formal or historical claim on paper.

Under the modern restrictive theory of sovereign (state) immunity, a foreign state remains immune for genuinely sovereign/governmental acts (acta jure imperii), but loses immunity for ordinary commercial activity (acta jure gestionis) — the commercial activity exception. Diplomatic immunity is a separate doctrine protecting accredited diplomats from the host state’s legal process.

Example. A foreign state-owned airline breaches an ordinary commercial parts-supply contract — likely not immune, since this is commercial activity, not a sovereign act. Now compare: that same foreign state is sued over a purely governmental decision, like denying a visa application — likely immune, since this is a genuinely sovereign/governmental act the restrictive theory still protects.

Sovereign immunity’s American doctrinal roots trace back to The Schooner Exchange v. McFaddon, 11 U.S. (7 Cranch) 116 (1812), Chief Justice John Marshall’s foundational opinion. A French warship, once a privately owned American vessel before Napoleon’s forces seized it, docked in Philadelphia, and its original American owners sued to reclaim it. Marshall held U.S. courts lacked jurisdiction over the foreign sovereign’s warship at all — a nation implicitly waives part of its own territorial jurisdiction when it allows a friendly foreign sovereign’s public vessels or representatives into its ports, as a matter of international comity between equal sovereigns. For over a century, American courts read Schooner Exchange as granting essentially absolute immunity to any foreign state. That absolute approach was later abandoned in favor of the restrictive theory this section describes, eventually codified in the real Foreign Sovereign Immunities Act of 1976, which remains the actual U.S. statute governing suits against foreign states today, including its own explicit commercial activity exception.

Jus cogens (peremptory norms) — like the prohibition on genocide — bind all states, and no treaty, consent, or persistent objection can override them. Erga omnes obligations run to the international community as a whole, meaning any state has a legal interest in their observance, not just directly injured states.

Example. Two states sign a treaty purporting to authorize genocide between them — void under VCLT Article 53, since it conflicts with a peremptory norm, regardless of how clearly both states consented. The prohibition on genocide is also widely recognized as an erga omnes obligation — meaning any state, not just one directly harmed, has a legal interest in its observance.

The very term “erga omnes” entered modern international law through one specific, real passage: the ICJ’s judgment in the Barcelona Traction, Light and Power Company, Limited Case (Belgium v. Spain), 1970 I.C.J. 3. The case itself concerned a Canadian-incorporated utility company operating in Spain, and Belgium’s attempt to bring a claim on behalf of the company’s largely Belgian shareholders after Spanish courts effectively liquidated it — the Court ultimately rejected Belgium’s standing, since international law generally protects a corporation’s own national state, not its shareholders’ separately. But the judgment’s lasting significance comes from a famous piece of dicta distinguishing two kinds of international obligations: those owed only to another particular state, and those “the obligations of a State towards the international community as a whole,” specifically naming the outlawing of genocide as an example, in which “all States can be held to have a legal interest in their protection.” That single passage is the real origin of the erga omnes concept this section relies on.