Public International Law Notes | B.A. LL.B. (Five Year Course) Semester 9 | Mumbai University | munotes

So States built a set of rules, slowly and mostly out of self-interest, about the things they could not manage alone: the sea, ambassadors, war, treaties, trade. That is international law, and it exists because the alternative is worse for everybody, including the strong.

Its nature: how it differs from the law you already know

A student meeting this subject has already studied contract, crime and procedure. Those are all vertical systems: there is a legislature above you making the law, a court above you applying it, and a police force above you enforcing it. You are the subject of the system, not a maker of it.

International law is horizontal. The same States that are bound by the rules are the ones that make them. There is no world parliament that can pass a law binding India without India's agreement. There is no court that a State can be dragged before against its will. There is no international police force.

That single structural difference explains almost every feature of the subject that beginners find strange, and it is worth setting out plainly.

Do not read that table as saying international law is a weaker version of real law. It is a different kind of system, built for parties who are equals. Whether it is law at all is a genuine and famous argument, and it has a chapter of its own: [Is International Law Really Law?].

Its origin

The ancient world

Rules between political communities are as old as political communities. Treaty texts survive from the ancient Near East, the earliest well-known one being between Egypt and the Hittites in about 1259 BC, dealing with peace, the return of fugitives and mutual defence. Ancient India had its own body of rules on envoys, on the conduct of war and on treaties, discussed in the Arthashastra and in the Dharmashastra literature; ancient Greece had rules between city-States about heralds, truces and the treatment of the dead; and Rome had the ius gentium, meaning the law of peoples, which was really the part of Roman law applied to foreigners.

None of that is international law in the modern sense, because none of it presupposed a world of equal sovereign States. It matters for a different reason: it shows that wherever separate communities have dealt with each other regularly, they have produced rules, which is the strongest practical answer to anyone who says such rules are unnatural.

The birth of the modern system: 1648

The modern subject begins in Europe, and the conventional date is 1648.

The Thirty Years War had wrecked central Europe. It ended with the Peace of Westphalia, a pair of treaties signed at Munster and Osnabruck. What made Westphalia a beginning rather than just another peace was what it settled: that each ruler was supreme within his own territory, that no outside authority, including the Pope or the Emperor, could dictate the religion or the internal government of another State, and that States dealt with one another as legal equals.

That is the Westphalian system, and the word is still used. Its three ideas, territorial sovereignty, legal equality of States, and non-intervention in internal affairs, are the ideas that Chapter I of the United Nations Charter restated three centuries later.

Grotius

Hugo Grotius, a Dutch jurist writing in the same century, is called the father of international law, and the title is deserved even though he invented neither the subject nor the phrase.

His great work, De Jure Belli ac Pacis, meaning On the Law of War and Peace, appeared in 1625, in the middle of that war. Three things in it changed the subject.

He argued that the law between nations rests on natural law, meaning principles of reason that hold good whether or not any ruler has commanded them, and famously that they would hold good even if there were no God. That freed the subject from theology at a moment when Europe was tearing itself apart over theology.

He treated the practice and agreement of nations as a second source alongside reason. That is the ancestor of custom and treaty as we now know them.

And he wrote about the sea. In an earlier work, Mare Liberum, meaning The Free Sea, published in 1609, he argued that the sea cannot be owned by any nation and must be open to all. That argument, made for the commercial convenience of the Dutch, became the freedom of the high seas, and you will meet it again in [The High Seas].

Grotius is not the only name of the period and an answer that mentions only him is thin. Francisco de Vitoria in sixteenth-century Spain argued that the peoples of the Americas had rights that Spain was bound to respect. Alberico Gentili, an Italian who taught at Oxford, wrote on the law of war before Grotius did and insisted that theologians keep out of it. Samuel Pufendorf carried the natural-law tradition forward, Cornelius van Bynkershoek and Emmerich de Vattel in the eighteenth century turned it towards the practice of States, and Bynkershoek's cannon-shot rule, that a State's control at sea extends as far as a cannon can fire, is where the old three-mile territorial sea came from.

From Westphalia to the First World War

Between 1648 and 1914 the subject grew mostly by accumulation of practice, with occasional bursts.

The Congress of Vienna in 1815, after Napoleon, produced the first modern rules on diplomatic ranks and on free navigation of international rivers. The nineteenth century produced the first permanent international organisations, which were technical rather than political: the International Telegraph Union in 1865 and the Universal Postal Union in 1874, both still alive inside the United Nations family. The Hague Peace Conferences of 1899 and 1907 produced conventions on the law of war and set up the Permanent Court of Arbitration, which still sits at the Peace Palace and still decides cases in this book, including the one in [The Bay of Bengal Maritime Boundary Arbitration].

This is also the period the subject is criticised for. International law of this era was European law, applied among European States and applied to everyone else as an instrument of empire. Doctrines of the time openly divided the world into civilised, barbarous and savage nations, and the phrase "civilised nations" survives in article 38 of the Statute of the International Court of Justice to this day, an embarrassment that is dealt with in [General Principles of Law Recognised by Civilised Nations].

The balance of power

The organising idea of the European order between Westphalia and 1914, and MU has asked for it, so it needs stating and placing.

What it meant. No single State should become strong enough to impose its will on the rest, and to that end the others would combine against whichever State grew too powerful. It was a doctrine of policy, not a rule of law: no treaty required it and no court applied it, and it operated by shifting alliances, by territorial compensation when one State gained, and, when those failed, by war.

How it relates to law. It supplied the conditions in which the classical law could work. Sovereign equality made sense in a system of several roughly comparable Powers; the rules on neutrality, on the acquisition of territory in [Acquisition of State Territory], and on treaties of guarantee all assumed that States would balance one another. And it explains why the classical law tolerated war: if the ultimate corrective is a coalition against the over-mighty, then war cannot be prohibited, and it was not.

Why it failed, and what replaced it. The balance did not prevent the First World War; it helped to produce it, by turning a local quarrel into a general one through the alliances that were supposed to keep the peace. The response was collective security, which is the opposite idea: instead of shifting coalitions responding to power, a permanent organisation determines that a breach has occurred and the whole membership responds. The League tried it and failed, and the Charter tried it again, with the machinery in [The Use of Force and Collective Security].

The comparison an examiner wants. Under a balance of power, force is lawful and is the regulator. Under collective security, force is prohibited by article 2(4) and the only lawful uses are self-defence under article 51 and action authorised by the Security Council. Whether the Charter system has in fact replaced the balance, or whether the veto has preserved it inside the Council, is a fair question and the honest answer is that both are true in part.

Its development in the twentieth century

The League of Nations

The First World War killed something in the order of seventeen million people, and the response was the first attempt at a permanent, general, political organisation of States: the League of Nations, created by the Covenant that formed Part I of the Treaty of Versailles in 1919, with its seat at Geneva.

The League tried to do three new things. It tried to make war a matter of common concern rather than a private right. It set up the Permanent Court of International Justice at The Hague in 1922, the first standing world court, whose decisions are still cited and several of which appear in this book. And it created the International Labour Organisation in the same year, which outlived it and has a chapter of its own: [The International Labour Organisation].

It failed, and the reasons are worth knowing because they explain the design of what replaced it. The United States never joined, despite its President having proposed it. Germany, the Soviet Union and Japan joined and left. Decisions of the Council and Assembly generally required unanimity, so any one member could block action. The Covenant did not prohibit war outright; it required delay and inquiry before resorting to it. And the League had no armed force and its members would not act. Japan invaded Manchuria in 1931, Italy invaded Abyssinia in 1935, and in each case the League condemned and did nothing effective.

The Kellogg-Briand Pact of 1928, formally the General Treaty for the Renunciation of War, belongs to the same decade and is often asked about. Sixty-odd States renounced war as an instrument of national policy. It was ignored within a few years, but it was not useless: it is the first general treaty prohibition of war, and it is the basis on which aggression was later treated as a crime.

The United Nations

The Second World War killed something in the order of sixty million people. The Charter of the United Nations was signed at San Francisco on 26 June 1945 and came into force on 24 October 1945.

The Charter is where this subject now begins in practice, and Module III of your syllabus is largely about it. Four of its innovations changed international law itself rather than merely creating an organisation.

A general prohibition on force. Article 2(4) requires all Members to refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any State. That is a far wider prohibition than the League's, and the exceptions to it are narrow: self-defence under article 51, and action authorised by the Security Council under Chapter VII. This is the subject of [The Use of Force and Collective Security].

A body that can act. The Security Council under Chapter VII can take decisions binding on all Members, including sanctions and the use of force. The League had nothing like it. The price was the veto, and that is dealt with in [The Security Council: Composition, Voting and the Veto].

Self-determination and decolonisation. The Charter mentions self-determination of peoples in article 1(2), and Chapters XI and XII created obligations towards non-self-governing and trust territories. Within thirty years the membership of the United Nations more than trebled, and the new members were not European. That, more than anything else, changed whose law international law was.

A court with a Statute annexed to the Charter. The International Court of Justice replaced the Permanent Court, and every Member of the United Nations is automatically a party to its Statute.

After 1945: the four directions of growth

The subject has grown in four directions since, and every one of them is somewhere in your syllabus.

It stopped being only about States. International organisations became subjects of international law in their own right, which the International Court decided in 1949 in the Reparation for Injuries advisory opinion, worked in [Subjects of International Law]. Individuals acquired rights that they can assert and duties for which they can be punished. This is [Subjects of International Law].

It was written down. The International Law Commission was set up by the General Assembly in 1947 to promote the progressive development and codification of international law, and the conventions it produced are the backbone of the subject: the law of treaties in 1969, diplomatic relations in 1961, the law of the sea in 1982, and the articles on State responsibility in 2001. Four of those are the instruments this book teaches.

It reached inside the State. Human rights, refugees, the environment, health and trade are all now regulated by international law, and all of them concern how a State treats people and things within its own territory. That would have been unthinkable to a Westphalian diplomat.

It acquired institutions with teeth in specialised fields. The World Trade Organization has a dispute settlement system whose rulings States mostly obey, and the International Tribunal for the Law of the Sea and the International Criminal Court decide cases in their own areas.

India and international law

Your syllabus expects you to know India's position, and MU's own printed course objectives say so. Three points belong here and the detail follows later.

India was a founding member of the United Nations, and in a curious way a member before it was independent: British India signed the Charter in 1945. India was also an original member of the League and of the International Labour Organisation.

India's Constitution takes an unusually clear position for a constitution of its date. Article 51 directs the State to endeavour to promote international peace and security, to maintain just and honourable relations between nations, to foster respect for international law and treaty obligations, and to encourage settlement of international disputes by arbitration. It is a Directive Principle, so it is not enforceable in a court by itself, but it is a direction to the whole State.

And India has been a maker of the modern law, not only a receiver of it. Indian delegations shaped the law of the sea negotiations, India is a registered pioneer investor in deep seabed mining under the regime discussed in [The Area, and the Parallel System of Mining], and India has appeared before the International Court and before arbitral tribunals, including in the two cases your syllabus names by name.

A worked example, to show the whole system at once

Take a concrete problem and watch every part of this chapter come into play.

A cargo ship registered in Panama, owned by a Greek company, crewed mostly by Indians, is carrying goods from Dubai to Kolkata. Fifteen nautical miles off the Indian coast its engine fails, it drifts, and oil leaks into the sea, damaging fishing grounds. Who can do what?

Sovereignty and territory decide the first question. Fifteen miles out is beyond India's twelve-mile territorial sea but inside its exclusive economic zone, so India does not have full sovereignty there but does have sovereign rights over the living resources and jurisdiction over the protection of the marine environment. Those zones are drawn in [The Exclusive Economic Zone].

Treaty supplies the rules: the United Nations Convention on the Law of the Sea 1982, to which India, Panama and Greece are all parties, and separately the pollution conventions.

Nationality decides who may complain on whose behalf. The ship has Panamanian nationality by registration; the crew have Indian nationality; the company has Greek nationality. Which State may bring a claim for which injury is the subject of [Treatment of Aliens and Diplomatic Protection], and the answer is not obvious.

State responsibility decides whether Panama or Greece has broken an obligation and what it owes, which is [State Responsibility].

Municipal law decides whether an Indian court can try anybody, and whether India's own Territorial Waters, Continental Shelf, Exclusive Economic Zone and Other Maritime Zones Act 1976 gives it that power, because an Indian judge applies the Indian Act and not the Convention directly. Why that is so is [International Law in Indian Courts].

Notice what did not happen. No world government stepped in. No international police boarded the ship. Every step runs through rules the States themselves made and through courts they themselves agreed to use.

What it does NOT mean

It is not the law of the United Nations. The United Nations is one institution created by one treaty. International law is older and larger, and most of it operates nowhere near the UN.

It is not private international law. Private international law, also called conflict of laws, is a part of each country's own municipal law. It decides which country's law an Indian court should apply to a contract signed in Dubai and which country's judgments it should enforce. It governs disputes between private parties. Public international law governs relations between States. Your syllabus lists textbooks on both, so read the title of a book before trusting it for this paper.

It is not the same as foreign policy. What a State ought to do politically and what it is legally obliged to do are different questions. Confusing them is the commonest fault in weak answers.

It is not a Western invention that India merely received. That was a fair description in 1900 and it is not one now. Decolonisation changed the membership, and with it the content: the permanent sovereignty of States over their natural resources, the common heritage of mankind, and special and differential treatment in trade law were all pressed by newly independent States, India among them.

Limits and criticism

Three criticisms are serious and an examiner rewards a candidate who states them fairly.

It is weak where it matters most. International law works well on postal rates, aviation, shipping and trade, and works worst on war and the use of force, which is exactly where its failure costs most.

It reflects power. The five permanent members of the Security Council hold a veto that no other State holds. That is written into the Charter, and it means the enforcement machinery cannot be used against them or their close allies.

Its origins are not innocent. For most of its history it was made by a small group of States and applied to the rest, and doctrines such as terra nullius, meaning land belonging to no one, were used to justify taking land that was plainly occupied. The subject has moved a long way from this, but a student who does not know the criticism cannot answer a question that raises it.

Quick revision

International law is the body of rules governing relations between States and other subjects such as international organisations.

It is horizontal, not vertical: the same States make it, are bound by it, and enforce it. No world legislature, no compulsory court, no police.

Origin: rules between communities are ancient, but the modern system dates from the Peace of Westphalia, 1648, which established territorial sovereignty, the legal equality of States and non-intervention.

Grotius, De Jure Belli ac Pacis (1625), is the father of the subject: natural law as its basis, the practice of nations as a second source, and the freedom of the seas in Mare Liberum (1609). Also Vitoria, Gentili, Pufendorf, Bynkershoek and Vattel.

League of Nations, 1919: first general political organisation, the Permanent Court of International Justice, the ILO. Failed because the United States stayed out, decisions needed unanimity, war was restricted rather than banned, and there was no force behind it.

United Nations Charter, signed 26 June 1945, in force 24 October 1945: the article 2(4) prohibition of force, a Security Council that can bind Members, self-determination, and the International Court of Justice.

Since 1945: new subjects beyond States, codification by the International Law Commission, expansion into human rights and the environment, and effective institutions in specialised fields.

India: a founding member of the United Nations; article 51 of the Constitution directs the State to foster respect for international law and treaty obligations.

Do not confuse public international law, which is between States, with private international law, which is a part of each country's own law about which law applies to a foreign-element dispute.

Test yourself

1. Define international law in one sentence, then say what makes it different from Indian law.

International law is the system of rules governing relations between States and other subjects of international law such as international organisations. It differs from Indian law in being horizontal: the States bound by it are the ones who make it, no court has compulsory jurisdiction over a State that has not consented, and there is no central police force. Indian law is vertical, made by a legislature above those bound, applied by courts they must attend, and enforced by the State.

2. Why is 1648 treated as the birth of modern international law?

Because the Peace of Westphalia ended the Thirty Years War on terms that established the three ideas the modern system is built on: each ruler is supreme within his own territory, no outside authority may dictate another State's internal affairs, and States deal with one another as legal equals. Rules between communities existed long before, but not among equal sovereign States.

3. Why is Grotius called the father of international law, and is the title fair?

He is called that because De Jure Belli ac Pacis (1625) grounded the law between nations in natural law rather than theology, treated the practice and agreement of nations as a second source, and, in Mare Liberum (1609), argued for the freedom of the seas. The title is only roughly fair: Vitoria and Gentili wrote before him, and Bynkershoek and Vattel did as much to turn the subject towards State practice. He is the father in the sense of giving the subject its first systematic statement.

4. Give three reasons the League of Nations failed, and say how the United Nations Charter answered each.

The United States never joined, so the League lacked the strongest power; the Charter did not solve this by design but the UN has had near-universal membership. Decisions generally required unanimity; the Charter replaced this with majority voting, subject to the veto of the five permanent members. The Covenant restricted rather than prohibited war and provided no effective enforcement; article 2(4) prohibits the threat or use of force generally, and Chapter VII lets the Security Council take binding measures including force.

5. A question asks you to "critically examine the nature of international law". What are the two halves of that answer?

The first half is descriptive: it is horizontal, made by treaty and custom, binding chiefly on States, with consensual courts and decentralised enforcement. The second half is critical: it is weakest precisely where it matters most, on the use of force; it entrenches the power of the five permanent members through the veto; and its historical origins are European and imperial, though decolonisation has changed both its membership and its content. A good answer gives both halves and does not pretend the criticisms are unanswerable or fatal.

6. Distinguish public international law from private international law.

Public international law governs relations between States and other subjects of international law, and its sources are treaty, custom and general principles. Private international law, or conflict of laws, is part of a country's own municipal law and governs disputes between private parties with a foreign element: which country's law applies, which court has jurisdiction, and whether a foreign judgment will be enforced. An Indian statute or an Indian precedent can settle a private international law question; it cannot settle a public international law question.

"Considered legally binding." Not "binding", but considered binding, by the States themselves. That is a consent-based, positivist view: the law is what States accept as law.

"Civilised States." The nineteenth-century division of the world into civilised, barbarous and savage nations, dealt with in [What International Law Is, and Where It Came From]. The phrase is indefensible now, and later editions of the same treatise dropped it.

"In their intercourse with each other." States only, and only in their external dealings. Nothing about individuals, nothing about organisations, and nothing about how a State treats its own people.

The other classical definitions

An examiner expects more than one, and the four below are the ones the standard reading list carries.

J.L. Brierly: the law of nations, or international law, may be defined as the body of rules and principles of action which are binding upon civilised States in their relations with one another. Note what Brierly changes: he says binding upon, not "considered binding by". That is a small phrase carrying a large claim, because it does not make the obligation depend on the State's own opinion.

Charles Fenwick: international law may be defined in broad terms as the body of general principles and specific rules which are binding upon the members of the international community in their mutual relations. Fenwick's phrase members of the international community is wider than "States", and it is the opening through which organisations later walked.

J.G. Starke, whose Introduction to International Law your syllabus lists in two editions, gives the definition that is closest to the modern one and is the safest to reproduce: international law may be defined as that body of law which is composed for its greater part of the principles and rules of conduct which States feel themselves bound to observe, and therefore commonly do observe, in their relations with each other, and which includes also the rules of law relating to the functioning of international institutions and organisations, their relations with each other and their relations with States and individuals, and certain rules of law relating to individuals and non-State entities so far as the rights and duties of such individuals and non-State entities are the concern of the international community.

Torsten Gihl: the body of rules of law which apply within the international community or society of States. Short, and it moves the emphasis from the parties to the community they form.

Starke's definition is long, and its length is the point. It is long because it has to name three things Oppenheim's did not: international institutions, individuals, and non-State entities.

Why the classical definitions broke down

Four developments made "States only" impossible to maintain, and each one is a chapter of this book.

International organisations acquired legal personality. In 1949 the International Court of Justice was asked whether the United Nations could bring an international claim for injury to its agent, and in the Reparation for Injuries advisory opinion, worked in [Subjects of International Law], it held that the Organization has a large measure of international personality and the capacity to operate on the international plane, though it is certainly not a super-State. A definition that mentions only States cannot accommodate a claimant that is not a State.

Individuals acquired duties. The Nuremberg and Tokyo tribunals after 1945 tried individuals for crimes under international law, and the principle that individuals can be punished directly by international law is now embodied in the Rome Statute of the International Criminal Court. If international law can hang a man, it plainly reaches him.

Individuals acquired rights. Human rights treaties give individuals rights against their own State, and several of them let an individual petition an international body. That is the exact opposite of the Westphalian arrangement in which how a State treated its own subjects was nobody else's business.

Some rules stopped depending on consent. A small class of rules, called jus cogens or peremptory norms, binds every State whether or not it agreed, and no treaty contrary to them is valid. The prohibitions on genocide, slavery, torture and aggression are the standard examples. Article 53 of the Vienna Convention on the Law of Treaties gives them their statutory form, and they are worked out in [Validity of Treaties, and Jus Cogens]. A definition built entirely on consent cannot explain them.

The modern definition

Putting those four developments into a sentence gives the definition to write in an exam:

International law is the body of rules and principles which States, international organisations and, in defined respects, individuals and other entities accept as binding in their mutual relations, derived from treaties, custom and general principles of law, and including a small core of peremptory norms which bind regardless of consent.

Every clause of that sentence is doing work. States, international organisations and individuals answers the who. Accept as binding in their mutual relations keeps the consensual character that is still true of most of the subject. Treaties, custom and general principles names the sources, which are the subject of [The Sources of International Law]. A small core of peremptory norms admits the exception without overstating it.

A table you can reproduce under time pressure

A worked example

Two years ago the government of State A, a party to the Genocide Convention and to the Charter, orders its army to expel an ethnic minority from a border province. Thousands die. A commander, Colonel B, gives the orders on the ground. An international organisation, the World Food Programme, has a warehouse in the province, and one of its staff, an Italian national, is shot at a checkpoint.

Take Oppenheim's definition and ask what it can explain. It can explain that State A has broken a treaty it accepted as binding, and that other State parties may complain. That is all. It cannot explain Colonel B's personal criminal liability, because individuals are not in the definition. It cannot explain the organisation's claim for its warehouse and its staff member, because organisations are not in it either. And if State A had never ratified the Genocide Convention, Oppenheim's consent-based definition would struggle to explain why the prohibition binds it anyway.

Now take Starke's, with jus cogens added. Colonel B is liable because international law imposes duties directly on individuals. The World Food Programme may bring a claim for the injury to its agent under the reasoning in the Reparation for Injuries advisory opinion, and separately the agent's national State may protect him, the competition between the two being reconciled case by case. And the prohibition of genocide binds State A regardless of ratification, because it is peremptory.

That is what a change of definition buys, and it is what an examiner is testing when the question says "critically examine".

What it does NOT mean

A definition is not a source. No case has ever been decided by quoting Oppenheim. Definitions organise the subject; article 38 of the Statute of the International Court of Justice supplies the law.

Saying individuals are subjects does not mean they are equal subjects. An individual cannot make a treaty, cannot sit in the General Assembly and cannot bring a case before the International Court, whose Statute in article 34 says only States may be parties in cases before the Court. The correct statement is that individuals have some rights and some duties directly under international law, not that they stand where States stand.

"Civilised nations" is not a live test. Where the phrase survives, in article 38(1)(c), it is read as meaning nothing more than the legal systems of the world generally. No State is excluded from international law today for want of civilisation, and any answer that repeats the phrase without noticing the problem loses more than it gains.

Limits and criticism

Every definition is circular to a degree. Oppenheim defines the law by what States consider binding, which assumes they already know which rules are law. Brierly avoids the circle by asserting the rules are binding, which is the very thing an objector denies. There is no way out of this by better drafting; it is the same difficulty that makes [Is International Law Really Law?] a real question.

Definitions lag behind practice. Oppenheim wrote before the League, before the United Nations, before human rights treaties and before the International Criminal Court. Starke's definition is capacious enough for all of them because it was written after them, which is less a virtue of the author than a fact about dates.

The subject may now be too varied to define usefully. Trade law, the law of the sea, human rights and the law of armed conflict have their own institutions, their own courts and their own habits, so much so that international lawyers argue about the fragmentation of the subject. A single sentence that covers all of them will do so only at a level of generality that tells a student very little.

Quick revision

The phrase international law was coined by Bentham, 1780. The older name is the law of nations.

Oppenheim, 1905: the body of customary and conventional rules considered legally binding by civilised States in their intercourse with each other. Two sources, consent-based, States only, and the offensive adjective.

Brierly: rules binding upon civilised States. Says binding upon, not considered binding by.

Fenwick: binding upon the members of the international community.

Starke: the fullest of the classical definitions, expressly covering international institutions, individuals and non-State entities.

Gihl: rules applying within the international community or society of States.

Four developments broke the States-only definitions: international organisations acquired personality (the Reparation for Injuries advisory opinion, 1949, in [Subjects of International Law]); individuals acquired duties (Nuremberg, the International Criminal Court); individuals acquired rights (human rights treaties); and jus cogens, in article 53 of the Vienna Convention, binds without consent.

Modern definition to write: rules and principles binding on States, international organisations and, in defined respects, individuals, derived from treaties, custom and general principles, and including a peremptory core that binds regardless of consent.

Test yourself

1. State Oppenheim's definition and identify three defects in it.

"Law of Nations or International Law is the name for the body of customary and conventional rules which are considered legally binding by civilised States in their intercourse with each other." Three defects: it names only two sources and omits general principles of law, which article 38(1)(c) of the Statute of the International Court of Justice recognises; it limits the subjects to States, which cannot accommodate international organisations or the direct duties international law imposes on individuals; and it rests obligation entirely on what States consider binding, which cannot explain peremptory norms that bind without consent. The reference to "civilised States" is a fourth defect and is now indefensible.

2. What is the difference between saying rules are "considered legally binding by States" and saying they are "binding upon States"?

The first makes the law depend on the opinion of the States bound, which is a positivist, consent-based account; if a State does not consider a rule binding, on that view it is not bound. The second asserts an obligation independent of that opinion, which is what is needed to explain jus cogens and, more generally, why a new State is bound by customary rules it never voted for. Oppenheim takes the first position, Brierly the second.

3. Which definition would you use to answer a question about the liability of an army officer for war crimes, and why?

Starke's, because it is the only classical definition that expressly includes rules of law relating to individuals and non-State entities so far as their rights and duties are the concern of the international community. Oppenheim's, Brierly's and Gihl's are limited to States and cannot reach an individual at all.

4. Why is the phrase "civilised nations" still in the Statute of the International Court of Justice, and how should a student handle it?

Article 38(1)(c) was drafted in 1920 for the Permanent Court and reproduced in 1945 without change, so the phrase survives by inertia rather than by intention. A student should state that it is understood today as meaning the legal systems of the world generally, that it excludes no State, and that it is a relic of the era when international law was made by a small group of European powers and applied to everyone else.

5. "Any definition of international law is either too narrow to be accurate or too wide to be useful." Discuss briefly.

There is force in it. Narrow definitions such as Oppenheim's are accurate about the core, relations between States, but cannot accommodate organisations, individuals or peremptory norms. Wide definitions such as Starke's cover all of those but do so at such a level of generality that they tell a beginner little about what the rules actually are. The reply is that a definition is not meant to do the work of the subject: it fixes who is bound and by what, and article 38 and the chapters that follow supply the content. A definition that gets the who and the what right is doing its job even if it is uninformative about the rest.

6. Distinguish the sources of international law from the definitions of it.

A definition states what the subject is and whom it binds. A source states where a rule comes from and how you prove it exists. Bentham, Oppenheim, Brierly, Fenwick, Starke and Gihl offer definitions. Article 38(1) of the Statute of the International Court of Justice lists the sources: international conventions, international custom, the general principles of law recognised by civilised nations, and, as subsidiary means, judicial decisions and the teachings of the most highly qualified publicists. An answer that recites definitions when asked for sources, or the reverse, answers a different question from the one set.

What honesty requires. The absence of a legislature is real. It means international law changes slowly and unevenly, and that new problems can wait decades for a rule.

2. There is no compulsory court

The objection. Article 34 of the Statute of the International Court of Justice allows only States to be parties, and article 36 makes jurisdiction depend on consent, whether by special agreement, by a compromissory clause in a treaty, or by a declaration under the optional clause. A State that has consented to none of these cannot be sued.

The answer. Consent is required to start a case, not to be bound by the law. A State that refuses to appear is still in breach if it has broken a rule; it has merely avoided an adjudication. And the picture is far less bleak than the objection suggests: the World Trade Organization has a standing dispute settlement system that has decided hundreds of disputes between States, the International Tribunal for the Law of the Sea and Annex VII tribunals decide law of the sea cases, including the two your syllabus names, and regional human rights courts give judgments against States as a matter of routine.

What honesty requires. Some of the most serious disputes never reach any court, and a State can withdraw its optional-clause declaration when it fears a case.

3. There is no police force

The objection. When a State breaks the law nothing happens to it, so the rule was not really a rule.

The answer. Something does happen, but it happens differently. Self-help is lawful within limits: an injured State may take countermeasures, meaning conduct that would otherwise be unlawful, to induce the wrongdoer to comply, and the limits on that are in [Circumstances Precluding Wrongfulness, and Reparation]. Collective measures exist: under article 41 the Security Council may impose measures short of force, including complete or partial interruption of economic relations and of rail, sea, air, postal, telegraphic and radio communication, and under article 42 it may authorise the use of force. Those are stronger sanctions than most municipal legal systems apply to anybody. And the ordinary sanction is reciprocity and reputation: a State that repudiates its treaties finds its treaties repudiated and its future agreements distrusted.

What honesty requires. Enforcement is selective, and the veto means it is selective in a way that favours five States. That is the strongest form of the objection and it should be conceded.

4. The rules are broken

The objection. Wars are fought in defiance of article 2(4), so the rule is not real.

The answer. This proves less than it looks. Murder is committed daily and nobody concludes that the law of murder is not law. What matters is not whether a rule is broken but how the breach is treated. States that use force almost never say they are entitled to break the law. They argue self-defence under article 51, or invitation by the government, or Security Council authorisation. The very fact that a State in breach constructs a legal justification is evidence that it regards the rule as binding; a State that thought there were no rule would not bother.

What honesty requires. The breaches in this field are of a scale and consequence unlike those in a domestic system, and pointing to murder rates is a debating answer rather than a complete one.

5. It is really politics

The objection. Decisions in the Security Council are political, so international law is politics with legal vocabulary.

The answer. Some of it is, and the Council is a political organ. But most of international law never goes near the Council. Aircraft land, letters arrive, ships pass, patents are respected, extradition requests are processed and treaties are performed, every day, according to rules whose content is technical and whose application is not seriously contested. The politically charged fraction of the subject is the visible fraction, which distorts the impression.

The jurisprudential answers

The objection is a claim about what law is, so the answers are claims about that too.

Hart

H.L.A. Hart answered Austin by rejecting the command model altogether. In The Concept of Law he argued that a developed legal system is a union of primary rules, which tell people what to do, and secondary rules, which are rules about the rules: a rule of recognition identifying what counts as law, rules of change for making new law, and rules of adjudication for settling disputes.

Hart's own view of international law was mixed. He accepted that it is law rather than morality, because its rules are argued about in legal terms and are not felt as merely moral. But he doubted whether it had a rule of recognition, and thought it resembled the primary rules of a simple society more than a developed system.

That doubt is now weaker than it was when he wrote. Article 38 of the Statute of the International Court of Justice looks very much like a rule of recognition: it tells you what counts as international law and where to find it. The Vienna Convention on the Law of Treaties is a set of rules of change: how a rule is made, amended and ended. And the courts and tribunals named above are rules of adjudication, imperfect but real.

Kelsen

Hans Kelsen put international law at the top rather than the bottom. In his pure theory, every legal norm derives its validity from a higher norm, up to a basic norm, the Grundnorm. Kelsen argued that the basic norm of the whole legal order is international, and identified it with the principle pacta sunt servanda, meaning agreements must be kept, or with the principle that States ought to behave as they customarily have behaved.

On that view municipal law derives its validity from international law, not the other way round. This is the strong form of monism and it is dealt with in [International Law and Municipal Law].

The Austinian answer on its own terms

Even accepting Austin's framework, the modern answer is that sanctions do exist. Article 41 measures are sanctions. Countermeasures are sanctions. Suspension or expulsion from an organisation under article 6 of the Charter is a sanction. They are decentralised and unevenly applied, but Austin's own test asked whether an evil is likely to be visited on disobedience, not whether it is certain.

The argument that settles it in practice

Leave theory aside and ask what States do.

Every State maintains a legal adviser's department in its foreign ministry. Every State, when accused of a breach, answers in legal terms rather than saying the accusation is meaningless. Every State enters into treaties in the expectation that they will be performed, and performs the overwhelming majority of its own. Every State pleads international law in its own courts when it suits, and complains when another State ignores it.

A rule that nobody regarded as binding would not be argued about in this way. That is the practical answer, and it is stronger than any of the theoretical ones, because it is a fact about behaviour rather than a claim about definitions.

Distinctions that carry marks

Comity is the one students most often confuse with law. Comity means practices States follow out of goodwill or convenience, without regarding themselves as legally bound. When a State stops, the other State may be annoyed but cannot say a legal wrong has been done. The distinction is exactly the distinction between practice and custom in [Custom]: whether the practice is accompanied by a belief that the law requires it.

A worked example

State X and State Y share a river. For sixty years State X has released water from its dam every March so that State Y's farmers can irrigate. This year State X keeps the water. State Y protests, calling it a breach of international law.

Ask the three questions this chapter has been building towards.

Is there a rule? That depends on whether the sixty-year practice was accompanied by a belief on both sides that it was legally required, or was a neighbourly courtesy. If there is a treaty, the question answers itself. If there is not, State Y must show custom, and the test is in [Custom].

If there is a rule, is it law even though no court can be seized? Yes. If neither State has consented to the jurisdiction of the International Court, State Y cannot sue, but State X is in breach and owes reparation. Absence of a forum is not absence of an obligation, which is the answer to the second objection above.

What can State Y actually do? Protest formally, which preserves its position and prevents its silence being read as acquiescence. Propose negotiation, good offices, mediation, conciliation or arbitration under article 33 of the Charter, which is [Pacific Settlement of Disputes]. Take proportionate countermeasures, such as suspending its own obligations to State X under the same arrangement. Raise the matter in a regional organisation or the General Assembly.

Notice that the answer to "is it law" made no practical difference to any of that. What made the difference was whether a rule existed and what remedies were available, which are legal questions in exactly the ordinary sense.

What it does NOT mean

Saying it is law does not mean it is effective. Effectiveness and legal quality are different. A statute that is widely ignored is still a statute.

Saying it is weak does not mean it is optional. A State that breaks a rule and escapes consequences has still broken the rule, and the breach can be relied on against it later, in a claim, in a negotiation, or before a tribunal it eventually accepts.

The absence of a sanction in a particular case does not prove the absence of a legal rule. Municipal systems are full of legal obligations with no practical sanction, including many constitutional obligations of the State.

Limits and criticism

The strongest version of the sceptical case is not Austin's. It is this: international law is law, but it is law that operates most reliably where least is at stake, and least reliably where most is. Its enforcement machinery is structurally unavailable against five States and their close allies, and no amount of jurisprudence about rules of recognition changes that.

The honest answer is not to deny it but to say what follows. The subject remains worth knowing and using because the fraction of international relations that is contested by force is small, because the law shapes even the arguments of those who break it, and because the alternative to an imperfect legal order between States is not a perfect one but none.

Quick revision

The objection is Austin's: law is the command of a sovereign backed by a sanction; international law has no sovereign, no command and no reliable sanction, so it is positive international morality, not law.

The five heads of objection: no legislature, no compulsory court, no police, the rules are broken, it is really politics.

Answers: consent is not a defect, and custom, jus cogens and binding Security Council decisions already go beyond it; consent is needed to start a case, not to be bound; sanctions exist as countermeasures, article 41 measures and article 42 force; a State in breach argues self-defence or authorisation, which shows it accepts the rule; and most of the subject is technical and uncontested.

Hart: law is a union of primary and secondary rules. Article 38 of the ICJ Statute functions as a rule of recognition, the Vienna Convention as rules of change, the courts as rules of adjudication.

Kelsen: the Grundnorm of the whole legal order is international, identified with pacta sunt servanda.

The practical answer: States employ legal advisers, answer accusations in legal terms, and perform the overwhelming majority of their obligations. Nobody argues that hard about a rule they do not regard as binding.

Distinguish law from morality and from comity. Comity is courtesy, and its breach gives no legal claim.

Test yourself

1. State Austin's objection precisely, and identify the definition of law it rests on.

Austin defined law as the command of a determinate sovereign, habitually obeyed by the bulk of society and itself obeying nobody, backed by a sanction. International law has no determinate superior above States, its rules arise from agreement and practice rather than command, and no centralised sanction enforces them. Austin therefore classified it as positive international morality rather than law properly so called. The objection rests entirely on the command theory; it disappears if that theory of law is rejected.

2. How does Hart's account change the answer?

Hart rejected the command theory and defined a developed legal system as a union of primary rules of obligation with secondary rules of recognition, change and adjudication. On that account international law is law rather than morality, because its rules are treated and argued about as legally obligatory. Hart himself doubted whether it had a rule of recognition, but article 38 of the Statute of the International Court of Justice performs that function, the Vienna Convention on the Law of Treaties supplies rules of change, and the standing courts and tribunals supply rules of adjudication.

3. "International law cannot be law because it is constantly broken." Answer this in three sentences.

Frequency of breach does not determine whether a rule is law, or the law of murder would not be law. What matters is how breaches are treated, and States that use force almost invariably plead a legal justification such as self-defence under article 51 or Security Council authorisation, which shows they accept the rule and dispute only its application. A State that believed there was no rule would not construct an argument under it.

4. Distinguish law from comity, and give the test.

Law is a rule regarded as legally obligatory, so its breach founds a claim of illegality and a right to reparation. Comity is a practice followed out of goodwill or convenience without any sense of legal obligation, so its discontinuance gives ground for complaint but not for a legal claim. The test is the same as the test that separates practice from custom: whether the conduct is accompanied by a belief that the law requires it.

5. What sanctions does international law actually have? Name four.

Countermeasures by the injured State, meaning conduct otherwise unlawful taken to induce compliance, subject to proportionality. Measures not involving force decided by the Security Council under article 41, including interruption of economic relations and of communications. Force authorised by the Security Council under article 42. And institutional sanctions such as suspension of the rights of membership or expulsion under articles 5 and 6 of the Charter, together with the everyday sanctions of reciprocity, retorsion and loss of reputation.

6. Give the strongest form of the sceptical case, and the best reply to it.

The strongest form is not Austin's but this: international law works most reliably where the stakes are lowest and least reliably where they are highest, and its collective enforcement machinery is structurally unavailable against the five permanent members and their allies because of the veto. The best reply concedes the point and adds three things: most international relations are not contested by force and are governed effectively; the law shapes the conduct and the arguments even of those who breach it; and the alternative to an imperfect legal order between sovereign States is not a better one but none at all.

What is right about it

It explains obligations that consent cannot. It explains why genocide is unlawful for a State that never ratified the Genocide Convention. It supplies a standard for criticising the law rather than merely describing it, and it is where the moral energy of the human rights movement came from.

What is wrong with it

It cannot say what the rules are. Reason, appealed to by different people in different centuries, has produced opposite answers. Natural law was used to justify the colonisation it was also used to condemn, and doctrines of a natural hierarchy of civilisations were defended in the same vocabulary.

It has no way of settling a dispute. If two States both appeal to reason and disagree, nothing in the theory tells you which is right.

It makes the actual practice of States irrelevant, which is a strange result for a subject in which the practice of States is most of the evidence.

Positivism

What it says

The positivist school holds that international law is binding because States have consented to it. Law is not found, it is made, and it is made by the will of the States that are bound.

The evidence is the practice: treaties are agreements, and custom is a general practice accepted as law. Both look like consent.

Consent theory in its simple form says a State is bound only by what it has agreed to, expressly by treaty or tacitly by participating in the formation of a custom.

The auto-limitation theory, associated with Georg Jellinek, refines it. A sovereign State cannot be subjected to a will outside its own, so obligation must come from the State limiting itself. The State's own will, self-restricted, is the source of the obligation.

Bynkershoek and, in the nineteenth century, jurists such as Triepel and Anzilotti, developed the positivist account. Anzilotti built the whole system on pacta sunt servanda, meaning agreements must be kept, treated as an axiom that needs no further justification.

What is right about it

It matches how the subject actually works. Treaties are negotiated, signed and ratified; reservations are entered and objected to; States argue about whether they participated in a practice. All of that is the language of consent, and any theory that cannot account for it is describing some other subject.

It also matches the sources: article 38 of the Statute of the International Court of Justice lists conventions and custom first, and both are consent-based on their face.

What is wrong with it

New States. A State that became independent in 1960 was bound at once by the customary law of the sea, of diplomatic relations and of State responsibility, none of which it had participated in making. Pure consent theory has no comfortable answer, and the usual one, that the new State impliedly consents by entering the international community, is a fiction that does the work by assumption.

Jus cogens. Article 53 makes some norms non-derogable by agreement. If consent were the whole basis, States could agree to anything.

Auto-limitation is circular. If the State's own will is the source of obligation, the State can withdraw its will and end the obligation. A promise that binds only so long as the promisor wishes to be bound is not a promise. Anzilotti's answer, that pacta sunt servanda is an axiom outside the system, concedes that at least one rule is not consensual, which is exactly the naturalist's point.

Custom does not really need consent. The persistent objector rule, which the Fisheries case supports and which is set out in [Custom], allows a State to escape a new rule by objecting from the start. But a State that simply says nothing is bound, and silence is not consent in any ordinary sense.

The modern theories

Kelsen and the pure theory

Hans Kelsen tried to remove morality and politics from the account altogether. Every norm derives validity from a higher norm, and the chain must end somewhere, in a Grundnorm, a basic norm that is presupposed rather than derived.

For Kelsen the Grundnorm of the whole legal order is international, and he identified it either with pacta sunt servanda or with the principle that States ought to behave as they have customarily behaved. Municipal legal orders then derive their validity from international law, which is the strong monist position taken up in [International Law and Municipal Law].

The strength of this is its rigour. Its weakness is that a presupposed basic norm explains obligation only in the sense of naming the place where the explanation stops.

The sociological school

Leon Duguit and Georges Scelle held that law arises from the facts of social life. Human beings and communities are interdependent, that interdependence generates rules, and the State is not a mystical sovereign but a set of officials who happen to act. Scelle's dedoublement fonctionnel, or role splitting, describes how State officials act simultaneously as organs of their own State and as organs of the international community, because there are no separate international organs to do it.

This explains something the other theories struggle with: why the law grows fastest where interdependence is greatest, which is why aviation, posts, shipping, telecommunications, health and trade are its most developed fields.

The doctrine of fundamental rights

An older theory, still asked about, holds that States have certain fundamental rights inherent in statehood, usually listed as independence, equality, self-preservation and jurisdiction over their territory, and that international law is derived from these. It is closely tied to naturalism.

Its defect is that it is derived from a fiction. States are not people, and self-preservation as a legal right was used for a century to justify almost anything, until article 2(4) of the Charter and the narrow exception in article 51 replaced it with a rule.

The consent-plus-community view

Most working international lawyers today hold something less tidy than any of the above: that consent explains the great bulk of international law, that a narrow core of peremptory norms binds regardless of consent, and that the community interest expressed through the United Nations and through obligations owed to all States has grown at the expense of pure bilateralism.

That is not a compromise for want of nerve. It is what article 38 plus article 53 of the Vienna Convention actually amount to when read together.

The doctrines you must be able to state

Three phrases recur in this module and every one of them is a doctrine, not merely a maxim.

Pacta sunt servanda, agreements must be kept. Article 26 of the Vienna Convention on the Law of Treaties puts it in statutory form: every treaty in force is binding upon the parties to it and must be performed by them in good faith. It is worked in [Rights and Duties of the Parties, and Third States].

Jus cogens, a peremptory norm. Article 53: a norm accepted and recognised by the international community of States as a whole as one from which no derogation is permitted and which can be modified only by a subsequent norm of the same character. A treaty conflicting with one is void. It is worked in [Validity of Treaties, and Jus Cogens].

Rebus sic stantibus, things standing thus. The doctrine that a fundamental change of circumstances may allow a party to withdraw from a treaty. Article 62 of the Vienna Convention states it in a deliberately narrow form, and it is worked in [Termination and Suspension of Treaties].

Distinctions table

A worked example

State P has never ratified any human rights treaty and has never voted for any resolution on torture. Its officials torture a detainee, a national of State Q. State Q protests that State P has violated international law. State P replies that it has consented to no such rule.

On pure consent theory State P has an argument, and it is the argument that shows the theory's limits.

On naturalist theory State P is bound because the prohibition of torture is part of a higher law that consent cannot reach. The difficulty is that this proves too much and too easily; the same reasoning has been used to support propositions that were merely the speaker's convictions.

On the modern view State P is bound and the reasoning is legal rather than moral. The prohibition of torture is a rule of customary international law, established by the general practice of States accepted as law, and State P did not object to it as it formed. It is further a peremptory norm within article 53, so no State can contract out of it even by treaty. State P's consent is therefore not required, and the argument fails.

Notice that the modern answer used the positivist machinery of custom and then reached beyond it. That is exactly the hybrid position described above, and it is the answer to write.

What it does NOT mean

These are not rival descriptions of different subjects. They are rival explanations of the same rules, and a good answer treats them as competing accounts of one thing.

Positivism does not mean the law is whatever powerful States say. It means the law is what States have accepted as law, which is a different and much more constrained claim.

Naturalism is not the same as morality. A naturalist claims the higher law is legally binding, not merely that it ought to be followed. Confusing the two collapses the distinction the last chapter was built on.

A "theory" is not a source. Nobody wins a case by proving Kelsen right. The sources are in [The Sources of International Law], and they are the same whichever theory explains them.

Limits and criticism

The whole debate has a suspicious quality: it is largely a debate about how to describe a practice that goes on regardless. States concluded treaties and observed customs for centuries without agreeing on why they were bound, and they still do.

The debate earns its place for one reason. When a genuinely novel question arises, whether a new State is bound by old custom, whether a treaty permitting a monstrous thing is valid, whether a rule can bind an objector, no source settles it and the answer has to come from an account of why States are bound at all. That is when these theories stop being ornamental.

Quick revision

The question is: why is a State bound?

Naturalism: because of a higher law found by reason. Vitoria, Suarez, Pufendorf, and Grotius in part. Explains jus cogens; cannot say what the rules are.

Positivism: because States have consented, expressly by treaty or tacitly by custom. Bynkershoek, Triepel, Anzilotti. Auto-limitation (Jellinek): the State limits its own will. Explains treaty practice; cannot explain new States, jus cogens, or its own circularity.

Kelsen: validity flows from a presupposed Grundnorm, which is international, identified with pacta sunt servanda.

Sociological (Duguit, Scelle): law arises from social interdependence; Scelle's role splitting has State officials acting as organs of the international community.

Fundamental rights doctrine: States have inherent rights of independence, equality, self-preservation and jurisdiction. Now largely displaced by article 2(4) and article 51 of the Charter.

The working modern position: consent explains most of it, a peremptory core binds regardless of consent, and community interest has grown at the expense of pure bilateralism.

Three doctrines to be able to state: pacta sunt servanda (VCLT article 26), jus cogens (article 53), rebus sic stantibus (article 62).

Test yourself

1. Explain the auto-limitation theory and give the objection to it.

Jellinek's auto-limitation theory holds that since a sovereign State cannot be subjected to any will outside its own, the binding force of international law must come from the State restricting its own will. The objection is circularity: if the obligation exists only because the State wills it, the State can un-will it, and an obligation that lasts only as long as the obligor wishes is not an obligation at all. Anzilotti's response, treating pacta sunt servanda as an axiom outside the system, concedes that at least one rule does not rest on consent.

2. Why do new States create difficulty for consent theory, and how is the difficulty answered?

A State that became independent after the main body of customary law had formed never participated in making it, yet is bound by it from the moment of independence. Consent theory answers by implied consent: by entering the international community and claiming its benefits, the new State accepts its rules. The answer is a fiction, since no actual act of consent occurs and no new State has ever been permitted to opt out. The honest position is that custom binds by community acceptance rather than individual agreement, subject only to the persistent objector rule, which is available to a State that objects as the rule forms and not to one that did not exist then.

3. State article 53 of the Vienna Convention and explain which theory it supports.

A treaty is void if, at the time of its conclusion, it conflicts with a peremptory norm of general international law, meaning a norm accepted and recognised by the international community of States as a whole as one from which no derogation is permitted and which can be modified only by a subsequent norm of the same character. It supports the naturalist side of the argument, since it places some rules beyond the reach of consent, though it does so in positivist language by making the community's acceptance the test.

4. What does Scelle's role splitting explain?

That international law has almost no organs of its own, so the officials of States perform double duty: a national judge applying a treaty, or a naval officer arresting a pirate on the high seas, acts both as an organ of his own State and as an organ of the international community. It explains how a decentralised system functions at all, and it is a better description of enforcement in this subject than any analogy with a police force.

5. "The theories of the basis of international law are of no practical use." Discuss.

There is force in the criticism for ordinary cases: a treaty dispute is decided by the Vienna Convention and a boundary dispute by the law of the sea, and no tribunal needs to choose between Grotius and Anzilotti. The criticism fails for hard cases. Whether a newly independent State is bound by custom it never made, whether a treaty permitting slavery is void, and whether a persistent objector escapes a rule are all questions that no source answers on its face, and each requires a view about why States are bound at all. The theories are not tools of daily practice; they are what is left when the sources run out.

6. Distinguish pacta sunt servanda, jus cogens and rebus sic stantibus in one sentence each.

Pacta sunt servanda, in article 26 of the Vienna Convention, is the rule that every treaty in force binds the parties and must be performed in good faith. Jus cogens, in article 53, is a norm from which no derogation is permitted, so that a treaty conflicting with it is void. Rebus sic stantibus, in article 62, is the narrow doctrine that a fundamental and unforeseen change in circumstances that formed an essential basis of consent may be invoked to terminate or withdraw from a treaty.

The General Assembly asked the Court.

The case

Facts. Reparation for Injuries Suffered in the Service of the United Nations, ICJ Reports 1949, page 174, was an advisory opinion given on 11 April 1949. By a resolution of 3 December 1948 the General Assembly asked whether, where an agent of the United Nations suffers injury in the performance of his duties in circumstances involving the responsibility of a State, the Organization has the capacity to bring an international claim against the responsible de jure or de facto government for the damage caused to the Organization and to the victim, and how such a claim is to be reconciled with the rights of the victim's national State.

Held. The Court answered the first question unanimously and the second by eleven votes to four: the Organization has the capacity to bring an international claim, whether or not the responsible State is a Member. The Charter does not say so expressly, so the Court asked what characteristics the Charter was intended to give the Organization. It found that the Charter conferred on it rights and obligations different from those of its Members and entrusted it with the maintenance of international peace and security. The Court concluded that the Organization possesses a large measure of international personality and the capacity to operate on the international plane, though it is certainly not a super-State. As to the claim for the victim, there is no bond of nationality between the Organization and its agents, so the analogy of diplomatic protection does not by itself supply the answer. But agents may be sent on important missions to disturbed parts of the world and must be assured of effective protection, and the powers necessary to that end must be regarded as a necessary implication arising from the Charter. The Organization therefore has the capacity to exercise functional protection of its agents, resting on the breach of an obligation owed to the Organization itself. On the third question, by ten votes to five, competition between functional protection and the national State's diplomatic protection is to be reconciled case by case and by agreement.

Why it matters. Three separate doctrines come out of this one opinion, and each is asked about on its own.

International personality of organisations. After 1949 it is settled that an international organisation can be a subject of international law. The personality is derived, because it comes from the treaty its members made, and it is functional, because it extends only so far as the organisation's purposes require. That is the difference between an organisation and a State, whose personality is original and general.

Objective personality. The Court held the capacity good against a non-member. Fifty States, it said, representing the vast majority of the members of the international community, had the power to bring into being an entity possessing objective international personality. So the United Nations is not merely a creature effective between its own members.

Implied powers. Powers not written into the constituent treaty may be implied where they are necessary for the performance of the organisation's functions. This is now the standard method of reading the constitutions of international organisations.

The Charter's own provisions

The Charter itself deals with personality in a curiously modest way. Article 104 provides that the Organization shall enjoy in the territory of each of its Members such legal capacity as may be necessary for the exercise of its functions and the fulfilment of its purposes. Article 105 gives it privileges and immunities on the same functional test.

Notice that both provisions are about capacity in the territory of Members, which is municipal legal personality. Neither says anything about the international plane. That is precisely why the Court had to reason from the Charter's structure rather than quote a provision, and it is why the case is a leading one.

Individuals

The classical position

Under the classical theory the individual was an object, not a subject. An injury to a national abroad was an injury to his State, and it was the State's claim to bring, to settle or to abandon. That doctrine survives, in a much modified form, in [Treatment of Aliens and Diplomatic Protection].

How that changed

Duties came first. The Nuremberg Tribunal after the Second World War convicted individuals of crimes against peace, war crimes and crimes against humanity, and rejected the defence that they had acted as organs of a State. The Rome Statute of the International Criminal Court, in force since 2002, now gives a standing court jurisdiction over individuals for genocide, crimes against humanity, war crimes and aggression. Piracy is the older example: any State may try a pirate, whoever he is and whatever his nationality, and that rule is now in article 105 of the Convention on the Law of the Sea, in [The High Seas].

Rights came next. Human rights treaties give individuals rights against their own State, and several give them a procedure. Under the Optional Protocol to the International Covenant on Civil and Political Rights an individual may complain to the Human Rights Committee against a State that has accepted the Protocol. Regional systems in Europe, the Americas and Africa go further and let individuals bring cases to a court.

How far it has gone, stated exactly

An answer that says "individuals are now subjects of international law" without qualification is too strong. The precise position is this.

Individuals have duties directly under international law, and can be punished for breaching them by international tribunals, without any State's law intervening.

Individuals have rights directly under some treaties, and under a smaller set can enforce them before an international body.

Individuals cannot make treaties, cannot participate in making custom, and cannot be parties before the International Court of Justice, because article 34 permits only States.

The correct formulation is that individuals are subjects of international law to a limited extent, in defined respects, and that the extent has grown steadily since 1945.

Other entities

Your syllabus does not name these, but an examiner asking about subjects expects a candidate to know they exist.

The Holy See and the Vatican City. The Holy See makes treaties, called concordats, sends and receives diplomatic representatives, and is a party to several conventions. Its personality is not in doubt even though the Vatican City is minute.

Insurgents and belligerents. A rebel movement that controls territory and conducts hostilities in an organised way may acquire a limited personality: the law of armed conflict applies to it and it may make agreements about the conduct of hostilities.

National liberation movements. In the decolonisation era several were granted observer status at the United Nations and treaty-making capacity for limited purposes.

Non-governmental organisations. They are not subjects. They are creatures of some national law, they cannot make treaties, and their standing at the United Nations rests on the consultative status that article 71 of the Charter allows the Economic and Social Council to grant. That is influence, not personality, and the distinction is the whole point of [Amnesty International and Non-Governmental Organisations].

Multinational corporations. Also not subjects, in the orthodox view. They contract with States, and investment treaties allow them to arbitrate against States, but the capacity comes from the treaty between States rather than from any personality of their own.

Distinctions table

A worked example

The United Nations sends a mission to State R, which is not a member of the United Nations. A mission vehicle is destroyed and Ms S, a Canadian national employed by the mission, is seriously injured, both through the fault of State R's police.

Can the United Nations claim for the vehicle? Yes. It has international personality, which includes the capacity to bring a claim for damage to itself, and the Reparation opinion held that this holds good against a non-member because the personality is objective.

Can the United Nations claim for Ms S? Yes, by way of functional protection, because she is its agent injured in the performance of her duties, and the claim rests on State R's breach of an obligation owed to the Organization.

Can Canada claim for Ms S? Yes, by way of diplomatic protection, because she is its national. Both claims exist, and the Court expressly declined to lay down a priority: the competition is resolved case by case and by agreement, which in practice means the Organization and the national State coordinate.

Can Ms S claim in her own name? Not before the International Court, since article 34 permits only States to be parties. She may have a remedy in State R's own courts, and a remedy before a human rights body if State R has accepted the relevant procedure. That gap is the practical meaning of saying an individual is a subject only to a limited extent.

What it does NOT mean

"Subject" does not mean "equal". Saying an organisation or an individual is a subject says only that it has some rights and duties of its own. It says nothing about how many.

Personality is not conferred by a label. An entity does not become a subject because a treaty calls it a legal person. The test is functional: what can it actually do?

Being affected by international law is not the same as being a subject of it. A shipping company is affected by the law of the sea every day and is not a subject of it.

An organisation's personality does not make it a State. The Court was explicit that the United Nations is not a super-State, and its personality is limited to what its functions require.

Limits and criticism

The orthodoxy is under pressure from two directions and a good answer notices both.

From below. Corporations conclude agreements with States that look very like treaties, and arbitrate against States under investment treaties. Insisting they have no personality is increasingly a statement about form rather than substance.

From the side. The functional theory of organisational personality is elastic. If powers may be implied whenever they are necessary for an organisation's purposes, and the organisation itself decides what its purposes require, the constituent treaty stops being much of a limit. That objection was made at the time of the Reparation opinion and has never been fully answered.

Quick revision

A subject has rights and duties under international law directly and can act on them. An object is merely something the law is about.

States are the primary subjects: original, general personality. Article 34 of the ICJ Statute: only States may be parties before the Court.

International organisations: Reparation for Injuries Suffered in the Service of the United Nations, ICJ Reports 1949, page 174. The United Nations has a large measure of international personality, though it is not a super-State; the personality is objective, good against non-members; and it may exercise functional protection of its agents. The opinion is also the source of the doctrine of implied powers. Charter articles 104 and 105 give only municipal capacity and immunities, which is why the Court had to reason from structure.

Individuals: duties directly (Nuremberg, the Rome Statute, piracy), rights directly under some treaties with a procedure under fewer, but no treaty-making capacity and no standing before the ICJ. Subjects to a limited extent.

Others: the Holy See, insurgents and belligerents, national liberation movements. NGOs are not subjects; article 71 consultative status is influence, not personality.

Test yourself

1. Define a subject of international law and give the test.

A subject of international law is an entity possessing international legal personality, meaning the capacity to have rights and duties under international law directly, to bring international claims to vindicate them, and in some cases to make treaties. The test is functional rather than formal: not what a treaty calls the entity, but what it can actually do on the international plane.

2. State the facts and holding of the Reparation for Injuries opinion, and name the three doctrines that come out of it.

The General Assembly asked whether the United Nations could bring an international claim against a State responsible for injury to its agent, and how such a claim would sit with the rights of the agent's national State. The Court held that the Organization has a large measure of international personality and the capacity to operate on the international plane, though it is not a super-State; that it may claim against non-members as well as members; and that it may exercise functional protection of its agents, resting on breach of an obligation owed to the Organization, with the competing claim of the national State to be reconciled case by case. The three doctrines are the international personality of organisations, objective personality opposable to non-members, and implied powers.

3. Distinguish functional protection from diplomatic protection.

Diplomatic protection is a State's claim for injury to its national, and it rests on the bond of nationality; the classic authority on that bond is Nottebohm, in [Nationality]. Functional protection is an organisation's claim for injury to its agent, and it rests not on nationality, of which there is none, but on the breach of an obligation owed to the organisation itself. Both may exist at once on the same facts, and the Reparation opinion declined to rank them.

4. "Individuals are now subjects of international law." How would you qualify that statement?

They are subjects to a limited extent and in defined respects. They bear duties directly, and international tribunals from Nuremberg to the International Criminal Court have punished them without any national law intervening; piracy is the older instance of the same idea. They hold rights directly under human rights treaties, and under a smaller number they may petition an international body. But they cannot make treaties, cannot participate in the formation of custom, and cannot be parties before the International Court of Justice, since article 34 of its Statute admits only States.

5. Why is an international organisation's personality described as derived and functional, and what follows from that?

Derived because it comes from the treaty its member States concluded, unlike a State's personality, which comes from statehood itself. Functional because it extends only so far as the organisation's purposes require, which is the same test the Charter uses in articles 104 and 105 for capacity and immunities. What follows is that different organisations have different capacities, that an organisation cannot act outside its purposes, and that the doctrine of implied powers, which fills gaps by reference to those purposes, is both the strength and the weakness of the theory.

6. Are non-governmental organisations subjects of international law? Give reasons.

No. An NGO is incorporated under some national law, has no capacity to make treaties, bears no international responsibility, and cannot bring an international claim. What it has is access: article 71 of the Charter permits the Economic and Social Council to make arrangements for consultation with non-governmental organisations, and many have consultative status and considerable practical influence on the making of treaties and the reporting of breaches. Influence is not personality, and a candidate who confuses the two will get the question on Amnesty International wrong.

Held. By eleven votes to three, Liechtenstein's claim was inadmissible. It is the bond of nationality between the State and the individual that alone confers the right of diplomatic protection. The conferment of nationality is within a State's domestic jurisdiction, but to exercise protection is to place oneself on the plane of international law, and acts done within domestic jurisdiction do not necessarily have international effect. The question was therefore not whether the naturalisation was valid under Liechtenstein law, which the Court declined to examine, but whether it could be relied on against Guatemala. Arbitrators and courts faced with competing nationalities have preferred the real and effective nationality, that which accords with the facts and rests on the stronger factual ties. Nottebohm's connection with Liechtenstein was tenuous and his connection with Guatemala long-standing and close, and the naturalisation had been sought not to associate himself with the population of Liechtenstein but to substitute the status of a neutral for that of a national of a belligerent. It could not therefore be invoked against Guatemala.

Why it matters. It is the source of the genuine link doctrine. A State may confer its nationality on whom it likes, but other States need not recognise a nationality that has no factual foundation, and a claim brought on such a nationality is inadmissible. The same idea reappears at sea in the requirement of a genuine link between a ship and its flag State, in [The High Seas].

The modes of acquiring nationality

Six modes are recognised, and every legal system uses some combination of the first two with the rest as supplements.

By birth

Two competing principles, and their Latin names are asked for.

Jus soli, the law of the soil: a person acquires the nationality of the territory in which he is born. Common in the Americas.

Jus sanguinis, the law of blood: a person acquires the nationality of his parents wherever he is born. Common in Europe and Asia.

Most States mix them, and India is a good example of a State that has moved from one to the other. Section 3 of the Citizenship Act 1955 gives citizenship by birth in India, but the conditions have been tightened three times: for those born on or after 26 January 1950 and before 1 July 1987, birth in India was enough; for those born on or after 1 July 1987 and before the commencement of the Citizenship (Amendment) Act 2003, either parent had to be a citizen at the time of the birth; and for those born on or after that commencement, which the Act's own footnotes date at 3 December 2004, either both parents must be citizens or one must be a citizen and the other must not be an illegal migrant at the time of the birth. Section 3(2) excludes two cases whatever the dates: where a parent has diplomatic immunity and is not an Indian citizen, and where a parent is an enemy alien and the birth occurs in a place under enemy occupation. That is a shift from jus soli towards jus sanguinis, and it is a good illustration for an answer.

By descent

Section 4 gives citizenship by descent to a person born outside India whose parent was a citizen at the time of the birth, subject to registration at an Indian consulate within a prescribed period. This is jus sanguinis in its pure form.

By naturalisation

The grant of nationality to a foreigner who applies and satisfies conditions, usually residence, good character, knowledge of a language and an intention to reside. Section 6 of the Citizenship Act, read with the Third Schedule, sets India's conditions.

Naturalisation is where the Nottebohm problem arises, because it is the mode most easily used to manufacture a nationality of convenience.

By registration

A simpler route than naturalisation, offered to defined classes with an existing connection. Section 5 covers, among others, persons of Indian origin ordinarily resident in India for the prescribed period, persons married to Indian citizens, and minor children of Indian citizens.

By marriage

In older laws a woman took her husband's nationality automatically on marriage. That is now rare and is regarded as discriminatory. Modern statutes, including India's, treat marriage as a ground for eligibility to apply, not as an automatic transfer.

By subjugation, cession or option

When territory changes hands, the population of the transferred territory generally acquires the nationality of the successor State, often with a right of option to keep the old one within a period. This connects to [State Succession].

The modes of losing nationality

By renunciation. Section 8 of the Citizenship Act allows a citizen of full age and capacity who is also a national of another country to make a declaration renouncing Indian citizenship, and minor children lose it with the parent, with a right to resume it within a year of attaining majority.

By termination. Section 9(1) is the provision to know for the next chapter: any citizen of India who by naturalisation, registration or otherwise voluntarily acquires the citizenship of another country ceases upon that acquisition to be a citizen of India. There is a proviso suspending the rule during a war in which India is engaged, until the Central Government directs otherwise. Section 9(2) leaves the determination of whether, when and how another citizenship was acquired to a prescribed authority.

By deprivation. Section 10 allows the Central Government to deprive a citizen of citizenship in defined cases, and it applies only to citizens by naturalisation or registration, or by virtue of certain constitutional provisions, and not to citizens by birth. The five grounds in section 10(2) are that the registration or certificate of naturalisation was obtained by fraud, false representation or concealment of a material fact; that the citizen has shown himself by act or speech to be disloyal or disaffected towards the Constitution; that he has, during a war in which India is engaged, unlawfully traded or communicated with an enemy or assisted one in business; that he has within five years after registration or naturalisation been sentenced in any country to imprisonment for not less than two years; and that he has been ordinarily resident outside India for a continuous period of seven years without, during that period, being a student abroad, in the service of a Government in India or of an international organisation of which India is a member, and without registering annually at an Indian consulate his intention to retain his citizenship.

Two safeguards sit on top of those grounds and are easy marks. Under section 10(3) the Central Government may not deprive a person unless it is satisfied that it is not conducive to the public good that he should continue to be a citizen. Under sections 10(4) to 10(6) it must give written notice of the ground, and on any ground except the seven-year residence ground the person may require the case to be referred to a Committee of Inquiry chaired by someone who has held judicial office for at least ten years, with two other members.

By long residence abroad, in some legal systems, and by substitution when a new nationality is acquired, which is what section 9 does.

Nationality of things that are not people

The word is used of ships, aircraft and corporations, and each has its own rule.

Ships have the nationality of the State whose flag they fly, and article 91 of the Convention on the Law of the Sea requires a genuine link between the State and the ship. This is [The High Seas].

Aircraft have the nationality of the State of registration, under article 17 of the Chicago Convention, in [International Air Law].

Corporations have, for most purposes, the nationality of the State of incorporation and of the registered office, and the question of who may protect a company and its shareholders is dealt with in [Treatment of Aliens and Diplomatic Protection].

A worked example

Mr T is born in Mumbai on 1 January 2006. His father is an Indian citizen and his mother holds a foreign passport but is lawfully resident in India. At twenty-two Mr T takes a job in State V and, after five years, accepts State V's offer of naturalisation. Two years later State V's police detain and beat him.

Was he an Indian citizen at birth? Yes. He was born after 3 December 2004, so section 3 required one parent to be a citizen, which the father was, and the other not to be an illegal migrant, which the mother was not.

What happened when he took State V's nationality? Section 9(1) applied automatically. By voluntarily acquiring the citizenship of another country he ceased upon that acquisition to be a citizen of India, and no order or hearing was needed.

Who can now protect him? State V, if anyone. India cannot, because he is no longer its national. But State V is the very State that injured him, and the general rule is that a State cannot exercise diplomatic protection against a State of which the injured person is also a national, still less against itself. In practice his remedy is State V's own courts and, if State V has accepted the procedure, a human rights body.

Would the answer change if State V's naturalisation had been a paper transaction with no residence and no ties? For India, no: section 9 turns on the fact of voluntary acquisition, not on its quality. For international law, possibly: on the Nottebohm reasoning another State could decline to recognise a nationality with no genuine link. The two systems can therefore give different answers to the same question, which is the point the chapter has been building towards.

Distinctions that carry marks

What it does NOT mean

A passport is not nationality. A passport is evidence of nationality and a request to allow the bearer to pass. It can be issued in error and it can be withheld from a national.

Domicile is not nationality. Domicile is a connection with a legal system for the purposes of private international law, chiefly succession and family law. A person can be an Indian national domiciled in England.

Nationality is not race, language or religion. Nationality here is a legal bond with a State. The word is used in ordinary speech for ethnic identity and that usage has no place in this answer.

The genuine link does not invalidate a nationality. Nottebohm did not hold that Nottebohm was not a national of Liechtenstein. It held that Liechtenstein could not use that nationality to bring a claim against Guatemala. The nationality remained good for every other purpose, including in Liechtenstein itself.

Limits and criticism

Nottebohm is narrower than students are taught. The Court dealt with a naturalisation obtained during a war for the express purpose of changing a belligerent status into a neutral one, and it repeatedly confined itself to the admissibility of Liechtenstein's claim against Guatemala. Read as a general requirement that every nationality must have a genuine link before any other State need respect it, the case would make millions of ordinary naturalised citizens vulnerable, and later practice, including the International Law Commission's work on diplomatic protection, has not read it that way.

The domestic-jurisdiction rule has costs. Because each State decides who its nationals are, and no State is obliged to grant nationality to anyone, statelessness can be created by legislative accident and nobody is obliged to cure it. That is the subject of the next chapter.

Investment citizenship strains the doctrine. Several States now sell nationality for a payment, with residence requirements that are nominal or absent. That is Nottebohm's problem in a modern form, and the law has not settled it.

Quick revision

Nationality is the legal bond between an individual and a State, and it is what entitles that State to exercise diplomatic protection for him.

Each State decides by its own law who its nationals are, but Nottebohm (ICJ Reports 1955, page 4) held that a nationality without a genuine link cannot be invoked against another State: the Court prefers the real and effective nationality, that which accords with the facts.

Modes of acquisition: birth (jus soli, place; jus sanguinis, blood), descent, naturalisation, registration, marriage, and subjugation or cession with a right of option.

Indian law: section 3 citizenship by birth, narrowed in 1987 and again on 3 December 2004; section 4 by descent; section 5 by registration; section 6 by naturalisation with the Third Schedule conditions; section 6A for the Assam Accord; section 6B inserted by Act 47 of 2019.

Modes of loss: section 8 renunciation, section 9(1) termination on voluntarily acquiring another citizenship, section 10 deprivation, which reaches only citizens by naturalisation or registration and never a citizen by birth.

Distinguish nationality from passport, from domicile and from ethnic identity.

Test yourself

1. Define nationality and say what turns on it in international law.

Nationality is the legal bond by which an individual is a member of a State and owes it allegiance. Three things turn on it: only the State of nationality may exercise diplomatic protection for the individual; a State may legislate for its nationals abroad under the nationality principle of jurisdiction; and a State must admit its own nationals and cannot expel them.

2. State the facts and holding of Nottebohm.

Nottebohm was a German national who lived and did business in Guatemala from 1905. In October 1939, a month after the war began, he obtained naturalisation in Liechtenstein with the residence requirement waived, and returned to Guatemala. Liechtenstein later claimed against Guatemala for its treatment of him. The Court held the claim inadmissible: conferring nationality is a matter of domestic jurisdiction, but exercising protection places the State on the international plane, and a nationality may be invoked against another State only where it reflects a genuine connection. His ties with Liechtenstein were tenuous and with Guatemala close, and the naturalisation was sought to exchange a belligerent status for a neutral one, so it could not be relied on against Guatemala.

3. Distinguish jus soli from jus sanguinis and say which India follows.

Jus soli confers nationality by place of birth; jus sanguinis confers it by descent from a national parent. India began with a broad jus soli in section 3 of the Citizenship Act 1955, then required one parent to be a citizen for births on or after 1 July 1987, and for births on or after 3 December 2004 requires one parent to be a citizen and the other not to be an illegal migrant. India therefore now applies a qualified jus soli heavily conditioned by descent, alongside a full jus sanguinis rule in section 4.

4. In what circumstances can an Indian citizen be deprived of citizenship, and who is protected from deprivation?

Section 10 of the Citizenship Act 1955 permits deprivation for obtaining citizenship by fraud, false representation or concealment of a material fact; disloyalty by act or speech towards the Constitution; unlawfully trading or communicating with an enemy in wartime; being sentenced in any country to imprisonment for not less than two years within five years after registration or naturalisation; and ordinary residence outside India for a continuous period of seven years without the exceptions and the annual registration that clause (e) requires. Section 10(3) adds that the Government must be satisfied that continued citizenship is not conducive to the public good, and sections 10(4) to 10(6) give a right to notice and, except on the residence ground, to a Committee of Inquiry. It applies only to citizens by naturalisation or registration and by virtue of certain constitutional provisions. A citizen by birth cannot be deprived under section 10 at all.

5. "The genuine link requirement makes a State's grant of nationality reviewable by other States." Is that right?

Not as stated. Nottebohm did not hold the naturalisation invalid, and the Court expressly declined to examine its validity under Liechtenstein law. What it held was that Liechtenstein could not invoke that nationality against Guatemala for the purpose of a claim. So other States are not reviewing the grant; they are declining to give it a particular international effect. Read more widely, the doctrine would unsettle the position of very large numbers of ordinary naturalised citizens, and later practice has not read it that way.

6. Mr A, an Indian citizen by birth, takes the citizenship of State B by naturalisation while working there. He does not tell anyone in India. What is his status?

He ceased to be an Indian citizen at the moment of acquisition, by force of section 9(1) of the Citizenship Act 1955, which operates automatically on voluntary acquisition of the citizenship of another country and requires no order. His silence is irrelevant to the result, though any question about whether, when or how he acquired the other citizenship falls to be determined under section 9(2) by the prescribed authority in the prescribed manner. India cannot thereafter exercise diplomatic protection for him, because he is no longer its national.

The problems double nationality creates

An examiner asks what is actually wrong with holding two passports. Five things.

Competing diplomatic protection. If A is a national of both State X and State Y and is injured by State Z, both X and Y may wish to claim, and Z faces two claimants for one injury.

Protection against the other State of nationality. This is the sharper problem. May State X claim against State Y on behalf of a person who is a national of both? The older rule said no, on the ground that the two claims cancel out. The modern rule, reflected in the International Law Commission's work on diplomatic protection, is that X may claim against Y where X's nationality is predominant, which is the Nottebohm idea applied to a different question.

Double military service. Both States may claim the person's compulsory service, and performing it for one may be treason to the other.

Double taxation and conflicting allegiance, resolved in practice by tax treaties rather than by nationality law.

Conflicting obligations in wartime. A dual national of two belligerents is in an impossible position, and the Nottebohm facts show what people do about it.

How international law handles them

The dominant and effective nationality

The working solution is not to deny that a person has two nationalities but to ask which one counts for the purpose in hand.

That is the doctrine of the dominant and effective nationality, and its clearest statement is the passage in the [Nationality] chapter from the Nottebohm case: where two States have conferred their nationality on the same individual and the situation extends beyond the domestic jurisdiction of either, arbitrators and courts have preferred the real and effective nationality, that which accords with the facts and rests on stronger factual ties.

The factors are habitual residence, the centre of family and business life, participation in public life, attachment shown for a country and inculcated in children, and the place where taxes are paid and duties performed.

The treaties

The Hague Convention on Certain Questions relating to the Conflict of Nationality Laws, 1930 is the classical instrument, and four of its articles are worth knowing by number.

Article 1 states the position this whole chapter turns on: it is for each State to determine under its own law who are its nationals, and this law shall be recognised by other States in so far as it is consistent with international conventions, international custom, and the principles of law generally recognised with regard to nationality. That second sentence is the treaty ancestor of the genuine-link reasoning in Nottebohm.

Article 3 accepts double nationality rather than denying it: a person having two or more nationalities may be regarded as its national by each of the States whose nationality he possesses.

Article 4 contains the rule about claims: a State may not afford diplomatic protection to one of its nationals against a State whose nationality such person also possesses.

Article 5 supplies the answer for everybody else: within a third State, a person having more than one nationality shall be treated as if he had only one, and that third State shall recognise exclusively either the nationality of the country in which he is habitually resident or the nationality of the country with which in the circumstances he appears to be in fact most closely connected. That is the dominant and effective nationality test, written into a treaty twenty-five years before Nottebohm.

The Convention relating to the Status of Stateless Persons, 1954 defines a stateless person as one who is not considered as a national by any State under the operation of its law, and gives such persons a status, identity papers and travel documents.

The Convention on the Reduction of Statelessness, 1961 obliges contracting States to grant nationality to a person born in their territory who would otherwise be stateless, and restricts deprivation of nationality where it would produce statelessness.

India is a party to none of the three. That is a fact worth stating in an answer rather than glossed over, and it is why the Indian discussion below runs on Indian statute alone.

India's position: no dual citizenship

India does not permit dual citizenship, and it says so twice.

Article 9 of the Constitution provides that no person shall be a citizen of India by virtue of article 5, or be deemed to be a citizen of India by virtue of article 6 or article 8, if he has voluntarily acquired the citizenship of any foreign State.

Section 9(1) of the Citizenship Act 1955 carries the rule forward beyond the commencement of the Constitution: any citizen of India who by naturalisation, registration or otherwise voluntarily acquires, or has at any time between 26 January 1950 and the commencement of the Act voluntarily acquired, the citizenship of another country shall upon that acquisition cease to be a citizen of India. There is a proviso: the sub-section does not apply to a citizen of India who during any war in which India may be engaged voluntarily acquires the citizenship of another country, until the Central Government otherwise directs. Section 9(2) leaves the question whether, when and how another citizenship was acquired to be determined by a prescribed authority in a prescribed manner and on prescribed rules of evidence.

Two points about the operation of section 9 are worth writing down.

It is automatic. The Indian citizenship ends at the moment of voluntary acquisition. No order is made and no hearing is held, which is what distinguishes it from deprivation under section 10, where an order, a notice and in most cases a Committee of Inquiry are required.

It turns on voluntariness. A nationality conferred by operation of another State's law without any act of the individual is not voluntarily acquired.

Overseas Citizenship of India

Because the pressure for dual citizenship from the Indian diaspora was real, Parliament created something that is deliberately not citizenship. Sections 7A to 7D of the Citizenship Act provide for registration as an Overseas Citizen of India Cardholder.

The name is misleading and the misdirection is the exam point. An OCI cardholder is a foreign national. The card gives a lifelong visa and certain parity with non-resident Indians in economic and educational matters, and it does not give the political rights that define citizenship: no right to vote, no eligibility for public office, no eligibility for public employment, and no right to acquire agricultural land. Section 7D allows the Central Government to cancel registration on stated grounds.

So India's answer to double nationality is to give the practical benefits people wanted while withholding the status, and a candidate who writes that India permits dual citizenship through the OCI card has made the error the provision was drafted to invite.

A worked example

Ms N is born in India in 1990 to Indian parents, so she is an Indian citizen by birth under section 3. She moves to State W for work at twenty-five, and at thirty-five she is naturalised there. She keeps her Indian passport and continues to use it when she visits her parents. At forty she is arrested and mistreated by the police of State Z while on holiday there.

Is she still an Indian citizen? No. Section 9(1) ended her Indian citizenship at the moment she was naturalised in State W, automatically. Continuing to hold and use an Indian passport does not preserve the status; the passport has simply become invalid, and using it knowingly is an offence.

Can India protect her against State Z? No, because she is no longer an Indian national and the right of diplomatic protection follows nationality.

Can State W protect her against State Z? Yes. She is its national and no competing nationality now exists.

Change the facts: suppose State W's law had made her a national at birth by descent through a grandparent, without any act of hers. Then she would be a dual national from birth, and section 9(1) would not have operated, because nothing was voluntarily acquired. India and State W could both wish to claim against State Z, and the question of which claim prevails, or whether both proceed, would be answered by asking which nationality is dominant and effective on the Nottebohm factors: habitual residence, the centre of her life, and the attachment she has shown.

Change them again: suppose it is State W itself that mistreats her. Now India would be claiming against the other State of her nationality. Under the 1930 Hague Convention's rule a State may not protect a national against a State whose nationality that person also possesses. The modern position is more flexible and permits the claim where the claimant State's nationality is predominant, and an answer that gives both positions and says which is modern will score better than one that gives only the older rule.

Distinctions table

What it does NOT mean

An OCI cardholder is not a citizen. No vote, no public office, no public employment, no agricultural land, and the registration can be cancelled under section 7D.

Statelessness is not the same as being a refugee. A refugee has a nationality and cannot or will not seek his State's protection because of a well-founded fear of persecution. A stateless person has no nationality at all. Many people are both, and the two categories are governed by different conventions.

Losing Indian citizenship under section 9 is not deprivation under section 10. Section 9 is automatic and applies to every citizen, including a citizen by birth. Section 10 requires an order and applies only to citizens by naturalisation or registration and by virtue of certain constitutional provisions.

Holding two passports does not by itself prove double nationality, and it usually proves that one of them should have been surrendered.

Limits and criticism

The world is moving the other way from India. A large and increasing number of States now tolerate dual nationality, on the view that it reflects how people actually live and that the old objections, chiefly military service and allegiance, have weakened. The Indian rule is a policy choice and can be defended, but an answer that presents it as the universal rule is wrong.

The 1961 Convention has few parties, so the strongest legal answer to statelessness binds a minority of States. The practical work is done by national law reform and by birth registration campaigns, not by international obligation.

The dominant nationality test is unpredictable. It asks a court to weigh residence, family, business and attachment with no fixed weights, which produces a defensible answer in a clear case and very little guidance in a close one.

Quick revision

Both problems come from one fact: every State writes its own nationality law and none is obliged to coordinate.

Double nationality arises from conflicting jus soli and jus sanguinis, from descent from parents of different nationalities, from naturalisation without renunciation, from marriage under older laws, and from transfers of territory.

Statelessness arises from the mirror gap, from deprivation, from failed acquisition, from State succession, and in practice most often from unregistered births.

The solution to double nationality is the dominant and effective nationality, the Nottebohm test: habitual residence, centre of family and business life, participation in public life, attachment shown.

Treaties: the Hague Convention on Certain Questions relating to the Conflict of Nationality Laws 1930, the Convention relating to the Status of Stateless Persons 1954, and the Convention on the Reduction of Statelessness 1961. India is a party to none of them.

India permits no dual citizenship: article 9 of the Constitution and section 9(1) of the Citizenship Act 1955, which operates automatically on voluntary acquisition. Sections 7A to 7D create the Overseas Citizen of India Cardholder, who is a foreign national with a lifelong visa and no political rights.

Test yourself

1. Explain how a person can come to hold two nationalities at birth, and why international law tolerates it.

Because States apply different principles. A child born in a State that follows jus soli, the law of the soil, to parents who are nationals of a State that follows jus sanguinis, the law of blood, acquires both nationalities at the moment of birth, and neither State has done anything irregular. International law tolerates it because the alternative would be to tell States whom they may treat as their nationals, which cuts against the settled rule that nationality is a matter for each State's own law.

2. May State A exercise diplomatic protection for a person who is a national of both A and B, against B?

Under the classical rule, no: article 4 of the 1930 Hague Convention provides that a State may not afford diplomatic protection to one of its nationals against a State whose nationality that person also possesses. The modern position, reflected in the International Law Commission's work on diplomatic protection, allows the claim where the claimant State's nationality is predominant, both at the date of injury and at the date of the claim. The test of predominance is the Nottebohm test of real and effective nationality.

3. Distinguish section 9 from section 10 of the Citizenship Act 1955.

Section 9 terminates citizenship automatically on the voluntary acquisition of the citizenship of another country. It needs no order, no notice and no hearing, and it applies to every citizen including a citizen by birth. Section 10 is deprivation by order of the Central Government on five stated grounds, it requires that the Government be satisfied that continued citizenship is not conducive to the public good, it requires written notice of the ground, and on every ground except long residence abroad the person may require a Committee of Inquiry. It applies only to citizens by naturalisation or registration and by virtue of certain constitutional provisions.

4. Is an Overseas Citizen of India a citizen of India? Give reasons.

No. Registration under sections 7A to 7D makes the holder a cardholder, not a citizen, and he remains a foreign national. The card confers a lifelong visa and parity with non-resident Indians in specified economic and educational matters. It confers none of the political incidents of citizenship: no right to vote, no eligibility for public office or public employment, and no right to acquire agricultural land. Section 7D permits cancellation of the registration. The name of the scheme invites the error, and the error is the point of the question.

5. What is statelessness, and what does international law do about it?

A stateless person is one who is not considered a national by any State under the operation of its law, which is the definition in the 1954 Convention relating to the Status of Stateless Persons. That Convention gives such persons a status, identity papers and travel documents. The 1961 Convention on the Reduction of Statelessness requires contracting States to confer nationality on a person born in their territory who would otherwise be stateless, and restricts deprivation that would cause statelessness. Both bind only their parties, and India is a party to neither.

6. "India's refusal to allow dual citizenship is out of step with international practice." Discuss briefly.

It is out of step in the descriptive sense: a large and growing number of States now tolerate dual nationality, on the view that it matches how people live and that the classical objections, chiefly competing military service and divided allegiance, have lost much of their force. India's position rests on article 9 of the Constitution and section 9(1) of the Citizenship Act and is a deliberate policy choice, softened in practice by the Overseas Citizen of India card, which supplies the economic benefits people sought while withholding political rights. Whether that is the right balance is arguable; what is not arguable is that the rule is a choice rather than a requirement of international law, which leaves nationality to each State.

Two rows of that table deserve a sentence each.

Nationality attaches to things that cannot vote. A ship has the nationality of its flag State, an aircraft of its State of registration, and a company of its State of incorporation. Nobody speaks of a citizen ship. That alone shows the two words are not synonyms.

A person can lose citizenship rights without losing nationality. A convicted prisoner may be disqualified from voting in many systems and remains a national throughout, so the State must still admit him and may still protect him abroad.

The position in Indian law

India uses the word citizenship in its Constitution and in its statute, and uses nationality almost nowhere. That is a drafting choice, not a denial of the distinction, and three features of Indian law show the distinction operating.

The Constitution deals with citizenship, and only at the commencement

Part II of the Constitution, articles 5 to 11, is headed Citizenship, and it is narrower than students expect. It determines who was a citizen at the commencement of the Constitution and then hands the subject to Parliament.

Article 5 makes a person a citizen at commencement if he had his domicile in India and was born in India, or either of whose parents was born in India, or who had been ordinarily resident in India for not less than five years immediately preceding the commencement.

Articles 6 and 7 deal with the migrations of Partition, in both directions.

Article 8 covers persons of Indian origin residing outside India, on registration with a diplomatic or consular representative.

Article 9, dealt with in [Double Nationality and Statelessness], excludes a person who has voluntarily acquired the citizenship of a foreign State.

Article 10 continues the rights of citizenship subject to any law Parliament may make.

Article 11 is the operative delegation: Parliament shall have the power to make any provision with respect to the acquisition and termination of citizenship and all other matters relating to citizenship. The Citizenship Act 1955 is made under it.

The Constitution then distinguishes citizens from persons

The distinction between a citizen and a mere national or foreigner does real work in Indian constitutional law, because some fundamental rights are given to citizens and others to persons.

Rights available only to citizens include those in article 15, prohibiting discrimination on grounds of religion, race, caste, sex or place of birth; article 16, equality of opportunity in public employment; article 19, the six freedoms including speech, assembly, association, movement, residence and profession; and article 29, protection of the interests of minorities.

Rights available to every person, citizen or not, include article 14, equality before the law, article 20, protection in respect of conviction for offences, article 21, protection of life and personal liberty, and article 25, freedom of conscience and religion.

That division is the constitutional expression of exactly the distinction this chapter is about: some things follow from being inside the community, and some follow from being human and present.

The Overseas Citizen of India: the clearest Indian illustration

Sections 7A to 7D of the Citizenship Act 1955 create the Overseas Citizen of India Cardholder, and the scheme is the best single illustration of the distinction available to an Indian student.

The cardholder is a foreign national. He holds another State's nationality and India's card. He gets a lifelong visa, and parity with non-resident Indians in specified economic, financial and educational fields. He does not get the vote, eligibility for public office, eligibility for public employment, or the right to acquire agricultural or plantation property. Section 7D permits cancellation of the registration.

So the OCI scheme hands over a package of practical benefits while withholding the political core. It is citizenship with the citizenship taken out, and the word in its name is the trap.

A worked example

Mr K holds the passport of State L. His grandfather was born in undivided India, so he qualifies for and obtains an OCI card. He lives in Mumbai for nine years, runs a business, pays Indian tax, and buys a flat.

Is he an Indian citizen? No. He is a national of State L, and his OCI card does not change that.

Can he vote in a Lok Sabha election? No. The vote is an incident of citizenship and the card confers none.

Can he be appointed to a post in the Union public service? No, for the same reason, and article 16's guarantee of equality of opportunity in public employment is in terms a right of citizens.

Can he buy an apartment? Yes, under the parity the scheme gives with non-resident Indians. Can he buy a farm? No; agricultural and plantation property is excluded.

If he is arrested in Mumbai, can he claim article 21? Yes. Article 21 protects every person, not only citizens. Can he claim article 19's freedom of movement? No; article 19 is a right of citizens.

If he is mistreated on a business trip to State M, which State can protect him? State L, because that is his State of nationality. India cannot, because he is not its national, and the card does not create the bond that diplomatic protection requires.

Notice how the same person is treated differently by the two systems, and how each answer followed from asking whether the question pointed inwards or outwards.

What it does NOT mean

Nationality is not the weaker status. It is the status that matters at the international level; citizenship is the status that matters domestically. Neither is a subset of the other in importance.

They are not always distinguished in drafting. Many statutes, and many States, use the two words as synonyms, and the United Kingdom's own legislation has used several categories of nationality with different citizenship consequences. A candidate should say the distinction is analytical and is not always reflected in the words a legislature chooses.

Citizenship is not the same as domicile or residence. Domicile is a private international law connection used for succession and family law; residence is a fact. A person can be an Indian citizen domiciled abroad, and a foreign national resident in India for decades.

An OCI cardholder is not an overseas citizen. The name says he is. The statute says he is a foreign national with a card. Trust the statute.

Limits and criticism

The distinction is blurring. In most modern States nationality and citizenship coincide exactly, because the class of nationals who are not citizens has all but disappeared with the empires that created it. Some writers say the distinction now survives mainly in textbooks.

But it has not disappeared. It survives wherever a State keeps a status that carries protection without political rights, and India has deliberately created such a status in the OCI card. It also survives structurally, because the word nationality still does work that citizenship cannot: it attaches to ships, aircraft and corporations, and it is the word that international tribunals use.

The gap it leaves is a real one. A person can be a lawful resident of a State for decades, pay its taxes, raise children who are its citizens, and have no vote in it, and this is a matter of policy that nationality law does not resolve.

Quick revision

Nationality points outwards and is the international law word: it entitles a State to protect the individual, obliges it to admit him, and follows him abroad.

Citizenship points inwards and is the municipal law word: it carries civil and political rights within the State.

Every citizen is a national; not every national is a citizen. The classical example is the colonial subject, protected abroad and without political rights at home.

Nationality attaches to ships, aircraft and corporations; citizenship does not.

Indian law: Part II of the Constitution, articles 5 to 11, fixes citizenship at commencement and article 11 hands the subject to Parliament, which enacted the Citizenship Act 1955.

Rights of citizens only: articles 15, 16, 19 and 29. Rights of every person: articles 14, 20, 21 and 25.

Sections 7A to 7D: the Overseas Citizen of India Cardholder is a foreign national with a lifelong visa and economic parity with non-resident Indians, and with no vote, no public office, no public employment and no agricultural land. Section 7D allows cancellation.

Test yourself

1. Distinguish nationality from citizenship, and give one example of each without the other.

Nationality is the status recognised by international law that links an individual to a State, entitling that State to exercise diplomatic protection and obliging it to admit him; citizenship is the municipal status carrying full civil and political rights within the State. Nationality without citizenship: the inhabitant of a colony who was a British subject, protected abroad but without political rights. Citizenship without nationality: this does not ordinarily arise for individuals, because citizenship presupposes membership of the State, but the point is illustrated in reverse by ships and aircraft, which have nationality and can have no citizenship at all.

2. Which fundamental rights in the Indian Constitution are available only to citizens?

Article 15, prohibition of discrimination on grounds of religion, race, caste, sex or place of birth; article 16, equality of opportunity in matters of public employment; article 19, the six freedoms; and article 29, protection of the interests of minorities. Articles 14, 20, 21 and 25 are available to every person, citizen or not.

3. Does the Constitution of India lay down a complete law of citizenship? Explain.

No. Part II, articles 5 to 11, determines who was a citizen at the commencement of the Constitution and deals with the migrations of Partition and with persons of Indian origin abroad. Article 9 excludes anyone who has voluntarily acquired a foreign citizenship, and article 10 continues the rights of citizenship subject to parliamentary law. Article 11 then expressly gives Parliament power to make any provision with respect to the acquisition and termination of citizenship and all other matters relating to it. The working law of citizenship is therefore the Citizenship Act 1955, made under article 11.

4. Explain the status of an Overseas Citizen of India, and why the name is misleading.

Registration under sections 7A to 7D of the Citizenship Act 1955 makes a person an Overseas Citizen of India Cardholder. He remains a national of some other State and is a foreign national in India. He receives a lifelong visa and parity with non-resident Indians in specified economic, financial and educational matters. He receives none of the political incidents of citizenship: no vote, no eligibility for public office or public employment, and no right to acquire agricultural or plantation property. Section 7D allows the registration to be cancelled. The name is misleading because the word citizen appears in a status that is expressly not citizenship, and India permits no dual citizenship at all.

5. Mr P, a national of State Q, has lived in India for twenty years on an OCI card. He is detained by the police without being told the grounds. Which constitutional rights can he invoke?

Article 21 and article 22 protections extend to every person, so he may complain of deprivation of personal liberty otherwise than according to procedure established by law, and of the failure to inform him of the grounds of arrest. Article 14, equality before the law, and article 20, protection in respect of conviction for offences, also extend to persons. He cannot invoke article 19, because the six freedoms are guaranteed to citizens, nor articles 15, 16 or 29. The division follows exactly the distinction between rights that attach to membership of the community and rights that attach to being a person within the State's power.

6. "The distinction between nationality and citizenship is of no practical importance today." Discuss.

There is something in it, because in most modern States the two classes coincide exactly, the colonial category of nationals without political rights having disappeared with the empires that created it. But the distinction has not lost its work. It survives in India in the Overseas Citizen of India scheme, which deliberately confers benefits without the political core. It survives structurally, because nationality attaches to ships under article 91 of the Convention on the Law of the Sea, to aircraft under article 17 of the Chicago Convention, and to corporations, and citizenship attaches to none of them. And it survives functionally, because international tribunals ask about nationality and never about citizenship: it was nationality, not citizenship, that decided Nottebohm.

"Recognized by civilized nations." The phrase is a relic of 1920 and is dealt with in [General Principles of Law Recognised by Civilised Nations].

"Subsidiary means for the determination of rules of law." Paragraph (d) is not a source of law at all. Judicial decisions and writings are ways of finding out what the law is, not ways of making it.

Formal and material sources

This is a distinction the standard textbooks make and examiners ask for.

A formal source is what gives a rule its legal validity, the process that makes it law. In this subject the formal sources are treaty and custom, with general principles a third.

A material source is where you find evidence of the rule's content: judgments, resolutions, treaty collections, diplomatic correspondence, the writings of jurists, the reports of the International Law Commission.

The distinction explains paragraph (d) exactly. Judicial decisions and the teachings of publicists are material sources: they tell you what the law is without making it.

Is there a hierarchy?

Students often assume article 38 lists the sources in order of authority. That is not quite right, and the exact position is worth stating.

There is no formal hierarchy in article 38. The order (a) to (d) is not stated to be an order of precedence, and the words "in the following order" were considered by the 1920 drafters and dropped.

There is a practical order. A court asked to decide a dispute will look first at any treaty binding both parties, because that is the most specific expression of what they agreed. Failing that, custom. Failing that, general principles, which in practice fill gaps rather than decide cases.

Paragraph (d) is expressly subordinate, being described as subsidiary means.

But three ordinary rules of legal reasoning apply, and they cut across the list. Lex specialis derogat legi generali: a special rule prevails over a general one, so a treaty between two States prevails over a general custom on the same point. Lex posterior derogat legi priori: the later rule prevails over the earlier, between the same parties. And jus cogens prevails over everything, so a treaty conflicting with a peremptory norm is void under article 53 of the Vienna Convention, whatever the parties agreed.

So the honest answer is: no hierarchy on the face of article 38, a working order in practice, and one true supremacy, that of peremptory norms.

The seven items MU lists, and where each is dealt with

MU prints topic 1.5 with a heading and then seven items, and each has a chapter.

Custom is [Custom]. The general practice and the belief that law requires it.

Treaties are [Treaties as a Source]. How a treaty binds parties, and how a treaty rule can pass into custom and bind non-parties.

General principles of law recognised by civilised nations are [General Principles of Law Recognised by Civilised Nations]. What the phrase means and which principles have actually been applied.

Judicial decisions and writings of jurists are [Judicial Decisions and the Writings of Jurists], and they are one chapter because article 38(d) treats them together as subsidiary means.

Equity is [Equity], including article 38(2)'s ex aequo et bono.

Resolutions of the General Assembly are [Resolutions of the General Assembly], which is the modern question article 38 does not mention at all.

What article 38 leaves out

This is the sharpest question on the topic and a good answer names four omissions.

Resolutions of international organisations. Article 38 was drafted before organisations produced resolutions in volume. General Assembly resolutions are not binding by themselves under articles 10 to 14 of the Charter, but they can express or crystallise custom, and some Security Council decisions bind all Members under article 25. Neither appears in article 38.

Unilateral acts. A State can bind itself by a public declaration intended to be binding, and no source in article 38 covers it.

Jus cogens. Article 38 is a list of where rules come from; it has nothing to say about a class of rules that outrank others. Article 53 of the Vienna Convention supplied that in 1969.

Soft law. Codes of conduct, guidelines, declarations and standards that are not binding but are widely followed. They are not law, and they are not nothing, and article 38 has no room for them.

A worked example

State A and State B dispute the boundary of their continental shelf. Where does a tribunal look, and in what order?

First, any treaty between them. If they have a delimitation treaty, that governs, because it is the most specific expression of their agreement, and lex specialis applies.

Second, any treaty in force between both on the subject generally. If both are parties to the Convention on the Law of the Sea, articles 76 and 83 govern, which is [The Continental Shelf].

Third, custom, if one of them is not a party. This is precisely the situation in the North Sea Continental Shelf cases, worked in [Delimitation of Maritime Boundaries], where Germany had signed but not ratified the 1958 Convention, so the tribunal had to ask whether the treaty rule had passed into custom.

Fourth, general principles, which here supply the vocabulary of equity but not the boundary.

And throughout, judicial decisions and writings, not as law but as evidence of what the law is. The tribunal will read the earlier delimitation judgments closely, and it is not bound by any of them, because article 59 provides that a decision of the Court has no binding force except between the parties and in respect of that particular case.

Distinctions table

What it does NOT mean

Article 38 is not a statute of the whole world. It is a direction to one court about what that court shall apply. Its authority as a list of sources comes from universal acceptance, not from its own terms.

A source is not the same as a basis of obligation. The basis of obligation is the question in [Theories of the Basis of International Law]. The source is the question of where a particular rule comes from and how to prove it.

Article 38(d) does not make judicial decisions binding precedents. Article 59 says a decision has no binding force except between the parties and for that case. The Court in practice follows itself closely, which is a different thing.

Ex aequo et bono is not equity. It is a power to decide outside the law altogether, and it needs the agreement of both parties. No party has ever given it. That distinction is worked in [Equity].

Limits and criticism

The list is a century old and shows it. It has nothing on resolutions, nothing on soft law, nothing on unilateral acts and nothing on peremptory norms, all of which are now part of any serious account of how international obligations arise.

"Civilized nations" should have been amended at some point in a hundred years and has not been.

The two main sources are becoming harder to tell apart. Multilateral conventions codify custom, custom then develops through the practice of parties to those conventions, and the two feed each other continuously. The North Sea cases had to unpick that relationship and it has only become more tangled since.

Quick revision

Article 38(1) of the Statute of the International Court of Justice is the accepted list: (a) international conventions, (b) international custom as evidence of a general practice accepted as law, (c) the general principles of law recognised by civilised nations, (d) subject to article 59, judicial decisions and the teachings of the most highly qualified publicists, as subsidiary means. Article 38(2): ex aequo et bono with the parties' agreement.

Its authority comes from universal acceptance and from article 93(1) of the Charter making every Member a party to the Statute.

The (b) wording is back to front: practice is evidence of custom, not the other way round.

Formal sources make law: treaty, custom, general principles. Material sources show what the law is: judgments, writings, resolutions, ILC reports.

No formal hierarchy, but a working order, plus lex specialis, lex posterior, and the true supremacy of jus cogens under article 53 of the Vienna Convention.

Omitted from article 38: resolutions of international organisations, unilateral acts, jus cogens, and soft law.

Article 59: a decision of the Court has no binding force except between the parties and in respect of that particular case.

Test yourself

1. Reproduce article 38(1) and identify which of its four heads is not a source of law.

The Court, whose function is to decide in accordance with international law such disputes as are submitted to it, shall apply: (a) international conventions, whether general or particular, establishing rules expressly recognized by the contesting states; (b) international custom, as evidence of a general practice accepted as law; (c) the general principles of law recognized by civilized nations; and (d) subject to the provisions of Article 59, judicial decisions and the teachings of the most highly qualified publicists of the various nations, as subsidiary means for the determination of rules of law. Head (d) is not a source: it is expressly a subsidiary means for determining rules of law, that is, a material rather than a formal source.

2. Is there a hierarchy among the sources?

Not on the face of article 38, which does not state an order of precedence and from whose draft the words indicating one were removed. There is a working order: a tribunal looks first to a treaty binding both parties, then to custom, then to general principles, and treats judicial decisions and writings as evidence throughout. Three rules cut across this: lex specialis, so a special rule prevails over a general one; lex posterior, so a later rule prevails over an earlier between the same parties; and the supremacy of jus cogens, which voids any treaty conflicting with it under article 53 of the Vienna Convention.

3. Distinguish formal from material sources with an example of each.

A formal source is the process that makes a rule legally valid: a treaty, or a custom formed by general practice accepted as law. A material source is evidence of the content of a rule: a judgment of the International Court, a report of the International Law Commission, or a treatise. A formal source can create an obligation; a material source can only show that one exists. Article 38(a) to (c) are formal, article 38(d) is material.

4. What does article 38 omit, and does the omission matter?

Resolutions of international organisations, unilateral acts of States, jus cogens, and soft law. It matters in the sense that any complete modern account of how international obligations arise has to deal with all four, and none of them can be fitted into the text without strain. It matters less than it seems, because resolutions and unilateral acts do their work largely through custom and through good faith, and jus cogens was given statutory form in article 53 of the Vienna Convention in 1969, which article 38 does not need to repeat.

5. Why is the wording of article 38(1)(b) criticised?

It says international custom is evidence of a general practice accepted as law, which reverses the logical relationship. Custom is the rule; the general practice accepted as law is the evidence from which the existence of the rule is inferred. Every court and writer reads the provision as though it said custom as evidenced by a general practice accepted as law, and the drafting has never been corrected.

6. State A obtains a judgment of the International Court against State B on a point of law. State C is later in dispute with State D on the same point. Is the earlier judgment binding?

No. Article 59 of the Statute provides that a decision of the Court has no binding force except between the parties and in respect of that particular case, and article 38(1)(d) is expressly made subject to article 59. There is no doctrine of stare decisis in international law. In practice the Court follows its own reasoning closely and departs from it only for good reason, so the earlier judgment will carry very great weight as a subsidiary means for the determination of the rule, but it binds nobody in the later dispute.

The psychological element: opinio juris

Opinio juris sive necessitatis means the belief that the conduct is required by law or by a sense of legal obligation.

It is what separates a customary rule from a courtesy. States have for centuries given a departing ambassador a ceremonial send-off; nobody thinks a legal obligation exists, so no custom has formed. States have for centuries refrained from arresting foreign diplomats; that is a rule, because States believe they are bound.

How it is proved. Opinio juris is a state of mind, and States do not have minds. In practice it is inferred from what States say when they act: from protests, from claims, from the terms of national legislation, from the reasons given for votes, and from the arguments States make when accused of a breach.

The paradox. For a new custom to form, some State must act first, and at the moment it acts there is no rule, so its belief that the law requires the act is false. The usual answer is that the belief may be prospective: the State asserts a rule and other States acquiesce, and the rule crystallises around the assertion. That is exactly what happened with the continental shelf, and the Truman Proclamation of 28 September 1945 is the standard example.

The leading case on how a treaty rule becomes custom

Facts. North Sea Continental Shelf Cases (Federal Republic of Germany/Denmark; Federal Republic of Germany/Netherlands), ICJ Reports 1969, page 3, decided 20 February 1969. Germany, Denmark and the Netherlands had agreed partial continental shelf boundaries near the coast in 1964 and 1965 but could not agree on their prolongation. Denmark and the Netherlands wanted the equidistance method of article 6 of the 1958 Geneva Convention on the Continental Shelf. Germany's North Sea coast is concave, and equidistance lines drawn from a recessing coast meet a short way out and cut the State behind them off from the shelf beyond. Germany had signed the 1958 Convention but never ratified it. By two special agreements of 20 February 1967 the parties asked the Court to state the applicable principles.

Held. By eleven votes to six, equidistance was not obligatory. Germany was not bound by article 6, having never ratified, and no estoppel arose against it. Equidistance was not inherent in the concept of the continental shelf, whose basis is the natural prolongation of the land territory, giving rights that exist ipso facto and ab initio. And article 6 had not passed into customary law. For that to happen the provision would have to be, at all events potentially, of a norm-creating character, and article 6 was not: it puts the obligation to use equidistance after a primary obligation to delimit by agreement, it is qualified by the notion of special circumstances whose meaning was controverted, and reservations to it were permitted, which the Court took as a legitimate inference that it was not thought to reflect emergent customary law. Further, a very widespread and representative participation in a convention might show that a conventional rule had become general law, but the number of ratifications and accessions was hardly sufficient. As to time, the passage of only a short period was not necessarily a bar, but within the period, however short, State practice, including that of States whose interests are specially affected, should have been both extensive and virtually uniform, and should have occurred in such a way as to show a general recognition that a rule of law was involved.

Why it matters. This is the leading statement of how a treaty rule becomes custom, and the three conditions it sets, norm-creating character, widespread and representative participation including specially affected States, and extensive and virtually uniform practice accompanied by opinio juris, are the answer to the standard essay question.

How a court finds a rule of custom: The Paquete Habana

The Paquete Habana; The Lola, 175 U.S. 677 (1900), Supreme Court of the United States, 5 January 1900.

Facts. Two fishing smacks, the sloop Paquete Habana and the schooner Lola, were regularly engaged in fishing on the coast of Cuba, sailing under the Spanish flag, each owned by a Spanish subject residing in Havana, with crews who also resided there and took shares of the catch. Each left Havana on a coast fishing voyage and was returning with a cargo of live fish when she was captured near Havana by a vessel of the United States blockading squadron during the war with Spain. Neither carried arms or ammunition; neither knew of the blockade or even of the war until stopped; neither ran the blockade or resisted; and there was no evidence that either vessel or crew was likely to aid the enemy. The District Court condemned both vessels and their cargoes as prize of war and they were sold.

Held, reversing the decrees, that the captures were unlawful and without probable cause, the proceeds of the sales to be restored to the claimants with damages and costs. Fuller CJ, with Harlan and McKenna JJ, dissented.

The method, which is why the case is in this chapter. International law is part of our law, and must be ascertained and administered by the courts of justice of appropriate jurisdiction, as often as questions of right depending upon it are duly presented for their determination. For that purpose, where there is no treaty, and no controlling executive or legislative act or judicial decision, resort must be had to the customs and usages of civilized nations; and, as evidence of these, to the works of jurists and commentators, not for the speculations of their authors concerning what the law ought to be, but for trustworthy evidence of what the law really is.

The rule found. Applying that method to several centuries of orders, treaties and practice, the Court held that at the present day, by the general consent of the civilized nations of the world, and independently of any express treaty or other public act, it is an established rule of international law that coast fishing vessels, with their implements and supplies, cargoes and crews, unarmed, and honestly pursuing their peaceful calling of catching and bringing in fresh fish, are exempt from capture as prize of war, and that prize courts administering the law of nations are bound to take judicial notice of that rule and to give effect to it in the absence of any treaty or other public act of their own government.

Three things to take from it.

It shows the two elements being proved. Centuries of practice, and orders and treaties showing that the practice was followed as a matter of obligation rather than of grace, which is opinio juris in the material a court can actually examine.

It fixes the place of the writings of publicists, and it is a better statement of article 38(1)(d) than the article itself: they are evidence of what the law is, not authority for what it should be. [Judicial Decisions and the Writings of Jurists] works that point.

It is the standard illustration of a municipal court applying custom directly, and it belongs with [International Law and Municipal Law] and with the Indian position in [International Law in Indian Courts], where the corresponding proposition comes from Gramophone Company of India Ltd v. Birendra Bahadur Pandey.

Local and regional custom

Custom need not be universal. A rule can bind a group of States, or even two.

Facts. Asylum Case (Colombia v. Peru), ICJ Reports 1950, page 266, decided 20 November 1950. On 3 October 1948 a military rebellion broke out in Peru and was suppressed the same day. Victor Raul Haya de la Torre, head of the American People's Revolutionary Alliance, was prosecuted for military rebellion. On 3 January 1949 the Colombian Ambassador in Lima granted him asylum, qualified him as a political offender and asked Peru for a safe-conduct. Peru refused, saying he had committed common crimes. The two States submitted the dispute to the Court.

Held. By fourteen votes to two, Colombia was not entitled to qualify unilaterally, and in a manner binding on Peru, the nature of the offence. The Court rejected each treaty Colombia relied on: the Bolivarian Agreement of 1911 recognised asylum in accordance with the principles of international law, and those principles do not entail a right of unilateral qualification; the Havana Convention of 1928 conferred no such right expressly or by implication; and the Montevideo Convention of 1933 had not been ratified by Peru. As to a regional American custom, Colombia had not proved the existence, regionally or locally, of a constant and uniform practice of unilateral qualification as a right of the State of refuge and an obligation on the territorial State, because the facts disclosed too much contradiction and fluctuation to disclose a usage peculiar to Latin America and accepted as law. By fifteen votes to one the Court also held Peru was not bound to grant a safe-conduct, and by fifteen to one rejected Peru's contention that Haya de la Torre was accused of common crimes, military rebellion not being in itself a common crime.

Why it matters. Two things. It establishes that a regional or local custom can exist but must be proved by the party asserting it, and to a strict standard: a constant and uniform practice accepted as law. And it is the leading case on diplomatic asylum, worked for that purpose in [Asylum].

The persistent objector

A State that objects to a rule from the beginning, and keeps objecting, is not bound by it when it crystallises.

The authority is the Fisheries case, worked in [Baselines and the Territorial Sea], where the Court held that the ten-mile rule for closing lines of bays had not acquired the authority of a general rule of international law and, in any event, was inapplicable as against Norway inasmuch as she had always opposed its application to the Norwegian coast.

Three limits on the doctrine are worth writing down.

The objection must be from the start. A State that acquiesces while a rule forms and objects afterwards is bound. Silence is treated as acquiescence.

It must be maintained. An objection abandoned is an objection lost.

It cannot work against jus cogens. No State can persistently object its way out of the prohibition of genocide.

Distinctions that carry marks

A worked example

For forty years the States bordering the Gulf of Aral have permitted each other's fishing vessels to shelter in their ports during storms without formality or charge. State E now begins charging State F's vessels a fee. State F protests that a customary rule has been broken.

Is there practice? Yes: forty years, consistent, and by all the States concerned. That is the easy half.

Is there opinio juris? That is the question, and State F must prove it. Evidence would include any statement by a government that the shelter was granted as of right, any legislation framing it as an obligation, any earlier occasion on which a State tried to charge and was told it could not, and any diplomatic protest. Evidence against would include statements describing the practice as a courtesy, or occasions on which shelter was refused without any complaint being made.

Whose custom would it be? Regional or local, binding only the Gulf States. On the Asylum reasoning State F must prove a constant and uniform practice accepted as law, and contradiction and fluctuation in the record will defeat it.

Could State E escape as a persistent objector? Only if it had objected as the rule formed and kept objecting. Forty years of granting shelter without protest is acquiescence, not objection.

What if there were a treaty? Then this whole analysis would be unnecessary, which is why a tribunal looks for a treaty first.

What it does NOT mean

Custom is not old practice. Age helps, but the North Sea judgment says a short period is not necessarily a bar. What matters is that the practice was extensive and virtually uniform and accompanied by the belief that law required it.

Custom is not unanimity. Generality is required, not universal participation. A State that says nothing is treated as acquiescing.

Breach of a customary rule does not destroy it. If the State in breach justifies itself by exceptions recognised within the rule, the rule is confirmed. It is only if States generally claim a right to act inconsistently that the rule erodes.

A treaty codifying custom does not abolish the custom. The two run in parallel, which is why non-parties remain bound by the customary rule and why a party that withdraws from the treaty remains bound by it.

Limits and criticism

Proving opinio juris is close to impossible in a hard case. Tribunals often infer it from the practice itself, which makes the two elements collapse into one and is circular. This is the most common criticism of the doctrine and it is a fair one.

It advantages the powerful. The practice of specially affected States counts for more, and States with navies, satellites and global reach are specially affected more often than others. Newly independent States made exactly this objection in the 1960s and 1970s, and it is why they preferred codification by treaty, where every State has one vote.

The relationship with treaties is now tangled. Conventions codify custom, parties then develop the custom through their practice under the convention, and non-parties are affected by rules they never joined. The North Sea judgment tried to keep the categories apart and the practice has run ahead of it.

Quick revision

Article 38(1)(b): international custom, as evidence of a general practice accepted as law. Two elements: general practice, and opinio juris sive necessitatis.

Practice includes acts, statements, legislation, national judgments, votes and treaties; it must be extensive and virtually uniform, including the practice of specially affected States; generality, not unanimity; no fixed duration.

Opinio juris is inferred from what States say when they act, and it is what separates custom from usage and comity.

North Sea Continental Shelf (ICJ Reports 1969, page 3): a treaty rule becomes custom only if the provision is of a norm-creating character, there is widespread and representative participation including specially affected States, and practice is extensive and virtually uniform with opinio juris. Article 6 of the 1958 Convention failed all three, partly because reservations to it were permitted.

Asylum (ICJ Reports 1950, page 266): local and regional custom exists but must be proved by the asserting party as a constant and uniform practice accepted as law; contradiction and fluctuation defeat it.

Persistent objector: Fisheries (ICJ Reports 1951, page 116). Norway had always opposed the ten-mile rule, so it was inapplicable to her. The objection must be from the start, maintained, and it cannot defeat jus cogens.

Test yourself

1. State the two elements of custom and explain why the second is needed.

The material element is a general and consistent practice of States; the psychological element is opinio juris sive necessitatis, the belief that the practice is required by law. The second is needed because practice alone cannot distinguish a legal rule from a habit. States have always shown courtesies to one another without regarding themselves as bound, and if practice alone made law every such courtesy would become an obligation and no State could ever stop.

2. What did the North Sea Continental Shelf cases decide about the passage of a treaty rule into custom?

That it can happen but requires three things. The provision must be, at all events potentially, of a norm-creating character, which article 6 was not, since it subordinated equidistance to prior agreement, was qualified by a controverted notion of special circumstances, and was open to reservation. There must be very widespread and representative participation in the convention, and the ratifications and accessions there were hardly sufficient. And State practice, including that of specially affected States, must have been extensive and virtually uniform, and must have occurred in a way showing general recognition that a rule of law was involved. A short passage of time is not by itself a bar.

3. Can a custom bind only two States? Give authority.

Yes. The Asylum case accepted in principle that a custom could be regional or even local, but held that Colombia had failed to prove one: she had not established a constant and uniform practice of unilateral qualification as a right of the State of refuge and an obligation upon the territorial State, because the facts disclosed too much contradiction and fluctuation to reveal a usage peculiar to Latin America and accepted as law. The burden lies on the State asserting the special custom and the standard is strict.

4. Explain the persistent objector rule and its three limits.

A State that has objected to a rule consistently from the time of its formation is not bound by it when it crystallises; the Fisheries case is the authority, the Court holding the ten-mile bay rule inapplicable to Norway because she had always opposed its application to her coast. The limits are that the objection must begin before the rule crystallises, since silence during formation counts as acquiescence; it must be maintained, since an abandoned objection is lost; and it cannot be raised against a peremptory norm, from which no derogation is permitted.

5. "Breach of a customary rule proves the rule does not exist." Is that right?

No, and the reasoning matters. If a State that departs from a rule seeks to justify itself by appealing to exceptions or justifications contained within the rule, its conduct confirms the rule rather than weakening it, because the State is not claiming a right to act inconsistently but arguing that its act falls within the rule. The rule erodes only where States generally claim a right to depart from it, and inconsistent conduct is then met without protest.

6. What is instant custom, and where is it said to have occurred?

The proposition that a customary rule can form very quickly where practice is extensive and virtually uniform and is accompanied by clear opinio juris, notwithstanding that little time has passed. The North Sea judgment leaves it open by saying that a short period is not necessarily a bar. It is usually illustrated by space law: within a few years of 1957 the practice of a very small number of space-faring States, with the acquiescence of everyone else, produced the rules that outer space is free for exploration by all and is not subject to national appropriation, which were then written into the Outer Space Treaty and are dealt with in [International Space Law].

Two consequences follow immediately, and both have their own provisions.

A party may not plead its own internal law as a reason for failing to perform. That is article 27, and its importance for an Indian student is dealt with in [International Law in Indian Courts].

A treaty binds only the parties. That is the next section.

Third States: the rule and its exceptions

The rule. Article 34 of the Vienna Convention: a treaty does not create either obligations or rights for a third State without its consent. The Latin tag is pacta tertiis nec nocent nec prosunt, agreements neither harm nor benefit third parties.

Obligations for a third State. Article 35 allows one, but on strict terms: the parties must intend the provision to create the obligation, and the third State must expressly accept it in writing.

Rights for a third State. Article 36 is easier: a right arises if the parties intend to accord it and the third State assents, and assent is presumed so long as the contrary is not indicated, unless the treaty provides otherwise.

Revocation. Article 37 provides that an obligation so created may be revoked only with the consent of the parties and of the third State, unless otherwise agreed; a right may generally be revoked by the parties alone, unless it was intended not to be revocable without the third State's consent.

And the exception that swallows the rule. Article 38 preserves the possibility that a rule set out in a treaty becomes binding on a third State as a rule of customary international law. That is the route that matters, and it is the next section.

How a treaty rule becomes binding on non-parties

This is the most heavily examined point in the topic, and the authority is the case worked in [Custom].

Facts. North Sea Continental Shelf Cases (Federal Republic of Germany/Denmark; Federal Republic of Germany/Netherlands), ICJ Reports 1969, page 3, decided 20 February 1969. Germany had signed but never ratified the 1958 Geneva Convention on the Continental Shelf. Denmark and the Netherlands, both parties, argued that the equidistance method in its article 6 bound Germany anyway, either because Germany had by its conduct assumed the obligations of the Convention or because article 6 had become a rule of customary international law.

Held. Neither argument succeeded. On conduct, only a very definite and very consistent course of conduct could justify holding a State bound where it had not carried out the formalities the convention itself laid down for becoming bound, and estoppel would require that Denmark or the Netherlands had detrimentally changed position in reliance, of which there was no evidence. On custom, the Court set out the three conditions: the treaty provision must be, potentially at least, of a norm-creating character; there must be very widespread and representative participation in the convention, including the States whose interests are specially affected; and State practice must have been extensive and virtually uniform in the sense of the provision invoked, and must have occurred in such a way as to show a general recognition that a rule of law or legal obligation was involved. Article 6 satisfied none of them.

Why it matters here. It is the answer to "can a treaty bind a non-party?". The answer is that the treaty cannot, but the rule in it can, if it has become custom, and the three conditions are the test.

The three relationships between a treaty and custom

Writers and tribunals describe three ways a convention and a customary rule can relate, and naming them is worth a mark.

Declaratory, where the treaty writes down a rule that was already custom. Much of the Vienna Convention on the Law of Treaties is declaratory, which is why its rules are applied to States that are not parties to it, and why India, which has not ratified it, is nevertheless governed by most of its content.

Crystallising, where a rule is emerging and the conclusion of the convention settles it.

Generating, where the convention creates a new rule that afterwards passes into custom through the practice of States, parties and non-parties alike. The exclusive economic zone is the standard example: invented in the negotiations that produced the 1982 Convention, and accepted as custom before that Convention entered into force.

Treaties and India

Three facts belong in an Indian answer and they are often missed.

India is not a party to the Vienna Convention on the Law of Treaties. It signed neither. It is nevertheless bound by most of the Convention's content, because those rules are declaratory of custom. So an answer that says "under the Vienna Convention India must..." is loose; the accurate form is that the rule binds India as custom, and the Convention states it.

The power to make treaties is executive, not legislative. Under article 73 of the Constitution the executive power of the Union extends to matters on which Parliament may make laws, and entries 13 and 14 of List I of the Seventh Schedule cover participation in international conferences and the entering into and implementing of treaties and agreements. So the Union executive may conclude a treaty without Parliament.

But implementing it may need legislation. Article 253 gives Parliament power to make any law for implementing any treaty, agreement or convention, notwithstanding the ordinary division of legislative powers. Whether a treaty is enforceable in an Indian court without such a law is the subject of [International Law in Indian Courts].

A worked example

State G, State H and thirty other States conclude the Convention on the Protection of Migratory Fish, which requires coastal States to close their fisheries for two months in the breeding season. State J, a major fishing State, does not sign. Ten years later, most States in the world, including many non-parties, have adopted closed seasons and defend them in the language of legal obligation. State J refuses.

Does the Convention bind State J? No. Article 34: a treaty does not create obligations for a third State without its consent, and State J has consented to nothing.

Has the rule become custom? Apply the North Sea test. Is the provision norm-creating? A flat obligation to close fisheries for a defined period looks norm-creating, unlike article 6 of the 1958 Convention with its priority for agreement and its special-circumstances escape. Is participation widespread and representative, including specially affected States? Thirty-two parties may not be enough by itself, but the relevant question is whether the great fishing States are among them. Has practice been extensive and virtually uniform with opinio juris? The facts say most States, including non-parties, have adopted closed seasons and defend them as legal obligations, which is exactly the evidence required.

Can State J escape as a persistent objector? Only if it objected from the start and has kept objecting. Ten years of silence followed by refusal is not enough, and this is where many answers go wrong.

What if State J had signed but not ratified? Then the North Sea reasoning applies directly: signature without ratification does not bind, and only very definite and consistent conduct, relied on to another State's detriment, could produce an estoppel.

Distinctions table

What it does NOT mean

A treaty is not superior to custom. Between the parties it prevails because it is more specific and usually later, not because treaties outrank custom in some hierarchy. Article 38 states no hierarchy.

Signing is not being bound. Signature usually authenticates the text and, under article 18 of the Vienna Convention, obliges the State not to defeat the treaty's object and purpose pending ratification. It does not make the State a party. Germany's position in the North Sea cases is the classic illustration.

A treaty codifying custom does not extinguish the custom. The customary rule continues in parallel, which is why non-parties remain bound and why a party that denounces the treaty is not released from the rule.

"Convention", "covenant", "protocol", "charter" and "pact" are all treaties. Nothing turns on the name.

Limits and criticism

The consent principle produces uneven law. The most powerful States can decline to join the conventions that would constrain them, and can then argue that the rules have not become custom because their own practice, as specially affected States, points the other way.

The North Sea test is demanding and vague at once. "Norm-creating character" and "extensive and virtually uniform" are not tests that decide a close case, and the Court's reliance on the permissibility of reservations as evidence that a provision was not thought to state emerging custom has been criticised as reading too much into a drafting choice.

Codification can freeze the law. Once a rule is written into a widely ratified convention, developing it through practice becomes harder, because departures now look like breaches of the treaty rather than the beginnings of a new custom.

Quick revision

Article 38(1)(a): international conventions, whether general or particular, establishing rules expressly recognised by the contesting States.

Law-making treaties lay down general rules; treaty contracts settle a particular matter. Only the first is a realistic candidate to pass into custom.

Article 26 VCLT, pacta sunt servanda: every treaty in force is binding on the parties and must be performed in good faith. Article 27: no plea of internal law.

Third States: article 34, no obligations or rights without consent (pacta tertiis nec nocent nec prosunt); article 35, an obligation needs express acceptance in writing; article 36, a right needs assent, which is presumed; article 37, revocation; article 38, the customary-law route is preserved.

North Sea Continental Shelf (ICJ Reports 1969, page 3): a treaty rule binds non-parties only as custom, and needs a norm-creating provision, widespread and representative participation including specially affected States, and extensive and virtually uniform practice with opinio juris. Signature without ratification does not bind, and estoppel requires detrimental reliance.

Three relationships between convention and custom: declaratory, crystallising, generating.

India has not ratified the VCLT but is bound by most of its content as custom. Treaty-making is executive under article 73 and List I entries 13 and 14; article 253 empowers Parliament to legislate to implement.

Test yourself

1. Distinguish law-making treaties from treaty contracts and say why the distinction matters.

A law-making treaty lays down general rules of conduct for a large number of States, such as the Charter or the Convention on the Law of the Sea. A treaty contract settles a particular matter between two or a few States, such as a boundary or an extradition arrangement. Article 38(1)(a) covers both, since it speaks of conventions whether general or particular. The distinction matters because only a law-making treaty contains a rule capable of general application, and so only a law-making treaty is a realistic candidate for passing into customary international law and binding non-parties.

2. Can a treaty bind a State that is not a party to it?

Not as a treaty. Article 34 of the Vienna Convention provides that a treaty does not create either obligations or rights for a third State without its consent, and article 35 requires express acceptance in writing before an obligation can arise. But article 38 preserves the possibility that a rule set out in a treaty becomes binding on a non-party as a rule of customary international law, and the North Sea Continental Shelf cases set the conditions: a norm-creating provision, widespread and representative participation including specially affected States, and extensive and virtually uniform practice accompanied by opinio juris.

3. Germany had signed the 1958 Convention. Why was that not enough?

Because signature is not the act by which a State becomes bound where the convention itself provides for ratification, and Germany never ratified. Denmark and the Netherlands argued that Germany's conduct had nonetheless assumed the obligations, but the Court held that where a convention specifically provides how the intention to be bound is to be manifested, it is not lightly to be presumed that a State which has not carried out those formalities has become bound another way. Only a very definite and consistent course of conduct, on which the other States had detrimentally relied so as to create an estoppel, would have sufficed, and there was no evidence of that.

4. Explain the three relationships a convention can have with custom, with an example of each.

Declaratory: the convention writes down a rule already customary, as much of the Vienna Convention on the Law of Treaties does, which is why its rules bind States that never ratified it, India among them. Crystallising: a rule is emerging and the conclusion of the convention settles it. Generating: the convention creates a new rule which afterwards becomes custom through subsequent practice, the standard example being the exclusive economic zone, which was accepted as customary before the 1982 Convention entered into force.

5. India has not ratified the Vienna Convention on the Law of Treaties. Is India bound by the rule that treaties must be performed in good faith?

Yes, but not by the Convention. Pacta sunt servanda is a rule of customary international law of great antiquity, and article 26 of the Convention is declaratory of it. So India is bound by the rule as custom, and the Convention is the best statement of its content. The correct way to put it in an answer is that the rule binds India as custom and the Convention states it, not that the Convention binds India.

6. Who in India has power to make a treaty, and can a treaty be enforced in an Indian court?

The power is executive. Article 73 of the Constitution extends the executive power of the Union to matters on which Parliament may legislate, and entries 13 and 14 of List I of the Seventh Schedule cover participation in international conferences and the entering into and implementing of treaties and agreements, so the Union executive may conclude a treaty without Parliament. Whether it can be enforced in a court is a separate question: article 253 empowers Parliament to legislate to implement any treaty notwithstanding the ordinary distribution of legislative powers, and the general position, worked out in [International Law in Indian Courts], is that a treaty affecting private rights or requiring a change in the law needs implementing legislation before a court will give effect to it.

The safer answer in an examination is the narrow view, with the acknowledgement that the Court has sometimes used the phrase loosely, and that the International Law Commission's recent work recognises both a category derived from national legal systems and a category formed within the international legal system itself.

Three limits operate whichever view is taken. The principle must be found in legal systems generally, not in one system. It must be transposable: rules built on institutions that have no international counterpart, such as a public prosecutor or a bankruptcy court, cannot cross over. And it applies only where treaty and custom do not, which is why paragraph (c) decides very few cases on its own.

The principles that have actually been applied

An answer that lists principles without authority is weak. These are the ones with a real pedigree.

Reparation for a breach. The proposition comes from the case worked below and is expressly framed as a principle of law in general, not merely of international law.

Facts. Factory at Chorzow (Germany v. Poland), Merits, PCIJ Series A, No. 17, judgment of 13 September 1928. Poland had taken measures of dispossession against two German-owned concerns at Chorzow in Upper Silesia, the owner and the operator of a nitrate factory. The Court had held in 1926 that this was not in conformity with the Geneva Convention concerning Upper Silesia of 22 May 1922. Negotiations on the consequences failed, and Germany began fresh proceedings in February 1927 claiming an indemnity.

Held. The obligation to make reparation exists, because it is a principle of international law, or even of law in general, that any breach of an engagement involves an obligation to make reparation. As to measure, restitution must in principle be in kind, or, if that is impossible, a sum corresponding to the value of the thing that cannot be restored, because reparation must as far as possible wipe out all the consequences of the illegal act and re-establish the situation which would, in all probability, have existed if that act had not been committed. Compensation for a seizure of property that could not lawfully be expropriated need not be limited to the value of the undertaking, and the Court ordered an expert enquiry to fix the value. It also held that although the loss fell on the companies, the claim was between States: reparation owed to a State does not change its character because a private person's loss measures it, and the rules of international law apply.

Why it matters. It is the classic instance of paragraph (c) at work, and both of its propositions are now in the ILC Articles on State Responsibility, worked in [State Responsibility] and [Circumstances Precluding Wrongfulness, and Reparation].

Good faith. Every system has it, and it appears throughout international law: in pacta sunt servanda in article 26 of the Vienna Convention, in the obligation to negotiate meaningfully stated in the North Sea cases, and in the interpretation rule in article 31.

Estoppel, sometimes called preclusion. A State that has by its conduct led another to rely on a state of affairs to its detriment cannot afterwards deny it. The North Sea cases discuss it and find it not established on the facts, which is itself instructive.

Res judicata. A matter finally decided cannot be reopened between the same parties. Article 60 of the Statute of the Court, providing that a judgment is final and without appeal, is its statutory expression.

Nobody may be a judge in his own cause, and the related rules of natural justice: notice, and an opportunity to be heard.

Circumstantial evidence and inference. The Corfu Channel case, worked in [Innocent Passage, and Straits], allowed a State that cannot get at evidence held inside the respondent's territory a more liberal recourse to inferences and circumstantial evidence, and that is a rule of proof taken from domestic systems.

Unjust enrichment, prescription, subrogation and abuse of rights appear in the case law and the literature, some of them more contested than others.

What paragraph (c) is not

It is not equity. Equity is dealt with in [Equity], and the difference matters: a general principle is a rule of law, applied because it is law. Equity in the sense of article 38(2), ex aequo et bono, is a power to decide outside the law and needs the parties' agreement.

It is not natural law. The principles are found by comparing legal systems, not by reasoning from first principles about what is right.

It is not a licence to invent. A judge who says "the general principles of law require X" without showing that X is found in legal systems generally has decided the case on his own view, and that is exactly the criticism made of this source.

A worked example

State M and State N conclude a treaty for a joint hydroelectric project. State N spends heavily on preparatory works after repeated written assurances by State M's minister that the project will proceed. State M then abandons it, and points out that the treaty contains no clause about withdrawal, that no custom governs the point, and that the Vienna Convention rules on termination do not cover a case like this.

There is no treaty rule, no custom, and no convention provision, and the tribunal cannot say the case is incapable of decision. This is where paragraph (c) does its work.

Good faith. The obligation to perform in good faith is common to every legal system and is expressly stated for treaties in article 26 of the Vienna Convention.

Estoppel. State M made clear representations, State N relied on them, and the reliance was detrimental. The North Sea cases show the elements a tribunal will look for and show that they must genuinely be established.

Reparation. If a wrong is found, the Chorzow principle supplies the remedy: as far as possible, wipe out all the consequences of the illegal act and re-establish the situation that would have existed. In practice that means compensation for the wasted preparatory works.

Notice that none of these three principles came from a treaty or from State practice. Each came from the shared content of domestic legal systems, and each was applied because it fits relations between States.

Distinctions table

What it does NOT mean

It is not a third-rank source. Article 38 states no hierarchy. Paragraph (c) is used less because treaty and custom usually answer the question first.

"Civilized nations" excludes nobody today. Any answer must say so.

Not every domestic rule crosses over. A principle tied to an institution international law does not have cannot be transposed, however universal it is at home.

Limits and criticism

It is the least predictable source. No agreed method exists for showing that a principle is general: no tribunal surveys all the world's legal systems, and in practice a handful of familiar ones stand in for the rest, which are usually the civil law and the common law of Europe.

It can conceal judicial law-making. A tribunal that wants a result and lacks a rule can reach for a "general principle", and there is no reliable check on it.

Its historical framing is still in the text. A century after the phrase was written, and eighty years after the Charter promised the sovereign equality of all its Members, the Statute still refers to civilised nations.

Quick revision

Article 38(1)(c): the general principles of law recognised by civilised nations.

Purpose: to prevent a non liquet, a declaration that the case cannot be decided for want of a rule.

"Civilized nations" is a 1920 relic, reproduced in 1945, never amended; read today as the legal systems of the world generally, excluding nobody.

Narrow view: principles drawn from municipal legal systems and transposable to relations between States. Broad view: also principles inherent in the international community. Three limits: generality, transposability, and subsidiarity to treaty and custom.

Principles actually applied: reparation for breach and the measure of it (Factory at Chorzow, PCIJ Series A No. 17); good faith; estoppel; res judicata; nobody a judge in his own cause; circumstantial evidence (Corfu Channel); unjust enrichment, prescription and abuse of rights.

Chorzow, the two sentences to remember: any breach of an engagement involves an obligation to make reparation; and reparation must as far as possible wipe out all the consequences of the illegal act and re-establish the situation which would, in all probability, have existed if that act had not been committed.

Distinguish general principles from equity and from natural law.

Test yourself

1. Why was article 38(1)(c) included in the Statute?

To prevent a non liquet. If a court could apply only treaty and custom it would have to dismiss for want of law any case falling outside both, and a court that regularly declared cases incapable of decision would not function as a court. Paragraph (c) gave the judges the shared content of the world's legal systems to draw on where the specific sources ran out.

2. What does "civilised nations" mean today, and how should a candidate handle the phrase?

Nothing more than the legal systems of the world generally. The phrase was written in 1920, when international law was made by a small group of European States, and it was reproduced without change in 1945 and never amended. A candidate should quote the provision accurately, state in a sentence that the adjective is a historical relic that excludes no State today, note that the International Law Commission has recommended dropping it, and then answer the question asked.

3. Give three general principles that have been applied, with authority for one.

Reparation for breach, good faith and estoppel. The authority for reparation is Factory at Chorzow, PCIJ Series A No. 17, where the Permanent Court held that it is a principle of international law, or even of law in general, that any breach of an engagement involves an obligation to make reparation, and that reparation must as far as possible wipe out all the consequences of the illegal act and re-establish the situation which would in all probability have existed had it not been committed.

4. Can any rule of Indian law be a general principle of law under article 38(1)(c)?

Only if two conditions are satisfied. It must be general, that is, found across the major legal systems of the world and not peculiar to India, since the source rests on what legal systems have in common. And it must be transposable to relations between sovereign States, which excludes rules built on institutions that have no international counterpart. A rule of the Code of Civil Procedure about the powers of a district court fails both tests; the principle of res judicata that underlies it satisfies both.

5. Distinguish a general principle of law from equity.

A general principle is a rule of law, applied because it is law, identified by comparing legal systems and imported where transposable. Equity, in the sense of article 38(2), is ex aequo et bono, a power to decide a case outside the law on grounds of fairness, which requires the express agreement of both parties and has never been given. Equity in a weaker sense, meaning fairness applied within the law, operates as part of the reasoning under the other sources, and is dealt with in [Equity].

6. "Article 38(1)(c) is a licence for judges to make law." Discuss.

There is force in it. No agreed method exists for establishing that a principle is general; no tribunal surveys the world's legal systems, and in practice a small number of familiar systems stand in for the rest. A tribunal that wants a result and has no rule can reach for a general principle and face little check. The answer is threefold: the source is genuinely necessary, because the alternative is a non liquet; the principles actually applied, reparation, good faith, estoppel, res judicata, natural justice, are uncontroversially common to legal systems everywhere; and paragraph (c) decides very few cases on its own, because treaty and custom answer the question first almost every time.

Which decisions count

The International Court of Justice and its predecessor, the Permanent Court of International Justice, whose judgments the present Court cites as its own line of authority.

Other international tribunals: the International Tribunal for the Law of the Sea, arbitral tribunals under Annex VII of the Convention on the Law of the Sea, including the two your syllabus names, the Permanent Court of Arbitration, the World Trade Organization panels and Appellate Body, and the international criminal tribunals.

Arbitral awards, which for much of the nineteenth and early twentieth centuries were where international law was actually made, and which remain important on territory and the sea.

National courts. This surprises students and it should not. A decision of the Supreme Court of India, or of any national court, applying international law is State practice as well as a subsidiary means, and so can help establish custom. That is a double function worth stating: a national judgment is evidence of what the law is, and is also part of the practice from which the law is made.

The writings of jurists

What they were

In the formative period they were the law's main repository. There was no world court, few reported arbitrations and no systematic record of State practice, so a court or a foreign ministry looking for the rule read Grotius, Bynkershoek, Vattel and their successors. Their authority was real and sometimes decisive.

What they are now

Their weight has fallen sharply, for the good reason that better evidence exists. Where there are judgments, treaty texts, published State practice and the reports of the International Law Commission, a court has no need to rely on a professor's opinion.

They still matter in three situations. Where the law is new and no case has arisen. Where practice is scarce or contradictory. And as a way of organising and stating what the sources show, which is what a court is doing when it says the position is settled and cites a treatise.

Whose writings

Article 38 says the most highly qualified publicists of the various nations, and the last three words are doing work: the writings drawn on should not all come from one legal tradition. In practice the Court rarely cites individual writers in its judgments, though individual judges do so freely in separate and dissenting opinions.

Collective writings carry more weight than individual ones, and this is the modern position that matters most. The International Law Commission, established by the General Assembly in 1947 to promote the progressive development of international law and its codification, is composed of persons of recognised competence and its draft articles and commentaries are treated with a respect no individual author now receives. The ILC Articles on State Responsibility of 2001 are the clearest example: not a treaty, not binding, and cited by the International Court as the statement of the law. The Institut de Droit International and the International Law Association occupy a similar, lesser place.

A worked example

State P and State Q are in dispute about whether State P may exercise jurisdiction over a foreign ship on the high seas in a situation no treaty covers. Counsel for State Q wants to argue that a customary rule prohibits it. What may she use, and how?

A judgment of the International Court on the same point in an earlier case between two other States. She may cite it, and it will carry great weight, but she must not say it binds State P: article 59 prevents that, and an opponent will say so. The correct submission is that the Court there stated the customary rule and that the rule, not the judgment, binds.

An arbitral award. The same, with somewhat less weight depending on the tribunal.

A decision of State P's own Supreme Court. This is the strongest material she has, and it does double duty: it is a subsidiary means for determining the rule, and it is State practice of the very State she is arguing against, which may also found an estoppel.

The ILC Articles on State Responsibility and their commentary, if the point concerns responsibility. Not binding, and treated as an authoritative statement.

A leading treatise. Usable, and weakest of all. If the treatise is the only support she has, she should expect the tribunal to notice.

Distinctions table

What it does NOT mean

Article 59 does not mean judgments are unimportant. It means they do not bind non-parties. The Court's jurisprudence is the single most valuable body of material in the subject.

A judgment is not a source even between the parties. It binds them because article 59 says so and because they submitted to the Court, and article 94(1) of the Charter obliges every Member to comply with the decision in any case to which it is a party. That is the force of a judgment, not the force of a rule of law.

"Publicists" does not mean journalists. It means writers on public international law.

The International Law Commission is not a legislature. Its draft articles become binding only if a convention is concluded on them and enters into force, or if their content is or becomes custom.

Limits and criticism

The absence of precedent is a real cost. States and their advisers cannot predict outcomes as confidently as in a domestic system, and the Court's practice of following itself is a convention rather than a rule.

The distribution of writers is uneven. The most cited publicists have overwhelmingly come from a small number of countries, so the phrase "of the various nations" describes an aspiration more than a practice.

The line between (d) and the other paragraphs is thinner than the Statute suggests. When the International Court states a customary rule in confident terms, that statement becomes the thing every later tribunal and every foreign ministry relies on, and the practical difference between finding the law and making it narrows.

Quick revision

Article 38(1)(d): subject to article 59, judicial decisions and the teachings of the most highly qualified publicists of the various nations, as subsidiary means for the determination of rules of law. Material sources, not formal ones.

Article 59: a decision of the Court has no binding force except between the parties and in respect of that particular case. No stare decisis, and the Court follows itself closely all the same.

Article 60: the judgment is final and without appeal. Article 94(1) of the Charter: every Member undertakes to comply with the decision in any case to which it is a party.

Article 62 intervention with permission; article 63 intervention as of right on the construction of a convention, with the construction equally binding on the intervener.

Decisions that count: the ICJ and the PCIJ; ITLOS, Annex VII tribunals, the PCA, WTO panels; arbitral awards; and national court decisions, which are both a subsidiary means and State practice.

Writings: decisive in the formative period, much weaker now. Collective writings, above all the International Law Commission, carry far more weight than individual authors.

Test yourself

1. Are judicial decisions a source of international law?

Not a formal one. Article 38(1)(d) describes them as subsidiary means for the determination of rules of law, and makes them subject to article 59, under which a decision of the Court has no binding force except between the parties and in respect of that particular case. So a judgment does not create a rule; it states one that already exists by treaty, custom or general principle. In practice the Court's judgments are the most authoritative statements of international law there are, and the Court follows its own reasoning unless there is reason to depart.

2. Explain article 59 and give two reasons for it.

It provides that the decision of the Court has no binding force except between the parties and in respect of that particular case. First, the Court's jurisdiction rests on consent, and it would be inconsistent to bind a State by a judgment in proceedings it was not party to and could not appear in. Second, it makes the intervention scheme coherent: article 62 lets a State with an interest of a legal nature request permission to intervene and article 63 gives a right of intervention where the construction of a convention to which it is a party is in question, in which case the construction is equally binding on it. If judgments bound everyone, those provisions would be pointless.

3. What weight does a decision of a national court carry?

Two kinds at once. It is a subsidiary means under article 38(1)(d) for determining what the rule is. It is also State practice of the State whose court gave it, and so can contribute to the formation and proof of a customary rule and may found an estoppel against that State. A national judgment against the interests of its own State is therefore particularly valuable evidence.

4. Why have the writings of jurists lost weight, and where do they still matter?

Because better evidence now exists. In the formative period there were no reported international judgments, few arbitrations and no systematic record of practice, so Grotius, Bynkershoek and Vattel were where the law was found. Today a tribunal has treaty texts, a large body of judgments, published State practice and the reports of the International Law Commission. Writings still matter where the law is new and no case has arisen, where practice is scarce or contradictory, and as an organised statement of what the other sources show. Collective writings, above all those of the International Law Commission, carry far more weight than individual authors.

5. The ILC Articles on State Responsibility are not a treaty. Why are they cited as if they were law?

Because much of their content is a statement of customary international law, and because their author is a body established by the General Assembly and composed of persons of recognised competence, which gives them an authority no individual writer commands. They are formally a subsidiary means under article 38(1)(d), and their binding force, where it exists, comes from the custom they state and not from the Commission's adoption of them.

6. "There is no precedent in international law." Is that a complete statement of the position?

No. It is accurate as a statement of formal doctrine, because article 59 excludes any binding force beyond the parties and the particular case, and article 38(1)(d) is expressly made subject to it. It is incomplete as a description of practice, because the Court cites its own judgments and those of the Permanent Court constantly, treats settled points as settled, and departs from its reasoning only for stated reasons. The accurate statement is that international law has no doctrine of binding precedent but does have a strong practice of consistency, and that the practical difference between the two is smaller than the doctrine suggests.

It is not the same as equitable principles. This is the distinction the whole chapter exists for. When the Court applies equitable principles in a maritime delimitation it is applying the law, because the law of delimitation itself requires an equitable result. It is not deciding ex aequo et bono, and it says so.

Where equity actually does its work: the sea

Maritime delimitation is where equity has been most important, and the two cases MU names are the authorities.

The North Sea Continental Shelf cases, ICJ Reports 1969, page 3, decided 20 February 1969, are worked in full in [Delimitation of Maritime Boundaries] and in [Custom]. On equity, what matters is the Court's conclusion: having rejected equidistance as neither a treaty obligation binding Germany nor a rule of custom, it held that the boundary lines were to be drawn by agreement between the parties and in accordance with equitable principles, taking account of the relevant circumstances, and that the parties were under an obligation to negotiate meaningfully rather than to go through a formal process. Equity there was not an escape from the law; it was the content of the applicable rule.

Continental Shelf (Tunisia/Libyan Arab Jamahiriya), ICJ Reports 1982, page 18, decided 24 February 1982, shows what that means in practice. By special agreement the parties expressly asked the Court to take account of equitable principles, the relevant circumstances characterising the area, and the new accepted trends at the Third UN Conference on the Law of the Sea. The Court found a single continental shelf common to both States, so physical criteria of natural prolongation could not determine what belonged to each, and it applied equitable principles to the relevant circumstances: the land frontier at Ras Ajdir, the conduct of the parties in granting petroleum concessions, the change of direction of the Tunisian coast at the Gulf of Gabes, and the Kerkennah Islands. It used proportionality between the shelf allotted and the length of the relevant coast as a test of equity, and described a two-segment line.

The point to take from both. The parties in Tunisia/Libya asked for equitable principles and did not confer power to decide ex aequo et bono. The Court decided according to law throughout. That is the clearest possible demonstration that equitable principles and ex aequo et bono are different things.

Article 83(1) of the Convention on the Law of the Sea now states the rule in treaty form: delimitation of the continental shelf between States with opposite or adjacent coasts shall be effected by agreement on the basis of international law in order to achieve an equitable solution. Article 74(1) says the same for the exclusive economic zone.

Other places equity appears

Equitable use of shared resources. International watercourse law is built on equitable and reasonable utilisation, which is an equitable standard rather than a mechanical rule.

The common heritage of mankind, in [The Common Heritage of Mankind], carries an explicitly equitable requirement: article 140 of the Convention on the Law of the Sea provides for the equitable sharing of financial and other economic benefits derived from activities in the Area.

Equitable geographical distribution of seats in international organs, which appears throughout the Charter, is a different use of the word: it is a political standard for composition and not a rule of decision.

A worked example

State R and State S have adjacent coasts. Neither has a delimitation treaty with the other. State R proposes a strict equidistance line. State S points out that its coast is concave, so equidistance would cut it off a short distance from shore, and asks the tribunal to do what is fair.

Can the tribunal simply do what seems fair? No, not in that sense. It cannot decide ex aequo et bono without the agreement of both parties, and State R has not agreed.

Must it then apply equidistance? No. The North Sea cases hold that equidistance is neither inherent in the concept of the shelf nor a rule of custom, and if both States are parties to the Convention on the Law of the Sea then articles 74 and 83 require delimitation by agreement on the basis of international law in order to achieve an equitable solution.

So what does it do? It applies equitable principles as the law requires, identifying the relevant circumstances of this area as the Court did in Tunisia/Libya: the general direction of the coasts, the concavity, any islands, the conduct of the parties, and any resource activity. It will check the result against proportionality between the areas allotted and the lengths of the relevant coasts.

What has the tribunal done, in terms of this chapter's classification? Equity infra legem. The applicable rule itself commands an equitable result, so applying equity is applying the law. Nothing has been set aside, and article 38(2) has not been touched.

Distinctions table

What it does NOT mean

Equity here is not the English law of equity. It has nothing to do with the Court of Chancery, trusts, specific performance or injunctions. The word means fairness.

Equitable principles are not ex aequo et bono. The Court said so and decided Tunisia/Libya on that footing. A candidate who writes that the Court decided the continental shelf cases ex aequo et bono has made the central error of the topic.

An equitable result is not an equal one. The Court in the North Sea cases expressly rejected the doctrine of the just and equitable share: its task was to delimit an area already appertaining to a State, not to apportion one, because the coastal State's rights exist ipso facto and ab initio.

Equity is not a source in article 38(1). It appears only in article 38(2), and its everyday operation is inside the other sources rather than alongside them.

Limits and criticism

It is unpredictable. A rule that says the result must be equitable, taking account of all the relevant circumstances, gives a tribunal very wide latitude and gives the parties' advisers very little to work with. The law of maritime delimitation has spent forty years trying to make the standard more predictable, which is why courts now use a structured three-stage method rather than an open weighing of circumstances.

It can conceal a decision made on other grounds. Equity requires no showing of practice or opinio juris, and a tribunal that wants a particular line can usually find circumstances to justify it.

Article 38(2) is a dead letter. A power never once exercised in more than a hundred years tells you something: States submit disputes to a court precisely because they want the law applied, and a State that would accept a decision on fairness alone would generally settle instead.

Quick revision

Three kinds. Infra legem, within the law: fair application and interpretation, needs no authority. Praeter legem, beside the law: gap-filling, overlapping with general principles. Contra legem: displacing a rule, permitted only under article 38(2) with the agreement of the parties.

Article 38(2): the Court may decide a case ex aequo et bono if the parties agree. Never used, before either the Permanent Court or the present Court.

Equitable principles are not ex aequo et bono. In North Sea Continental Shelf (ICJ Reports 1969, page 3) delimitation was to be by agreement in accordance with equitable principles and the relevant circumstances, with an obligation to negotiate meaningfully. In Continental Shelf (Tunisia/Libyan Arab Jamahiriya) (ICJ Reports 1982, page 18) the parties expressly asked for equitable principles and did not confer power to decide ex aequo et bono; the Court weighed the relevant circumstances and used proportionality as a test.

UNCLOS articles 74(1) and 83(1): delimitation by agreement on the basis of international law in order to achieve an equitable solution.

An equitable result is not an equal share: the North Sea cases rejected the just-and-equitable-share doctrine because shelf rights exist ipso facto and ab initio.

Test yourself

1. Distinguish equity infra legem, praeter legem and contra legem.

Infra legem operates within the law: the tribunal applies existing rules and uses fairness to choose among permissible constructions and to apply the rule to the facts. It needs no special authority. Praeter legem operates beside the law, filling gaps the rules leave, and overlaps with the general principles of law. Contra legem would displace a rule because applying it would be unfair, and a tribunal may do that only under article 38(2) of the Statute, which requires the agreement of both parties.

2. What is ex aequo et bono, and how often has it been used?

It is the power in article 38(2) to decide a case according to what is equitable and good rather than according to the existing rules, conferred only if the parties agree. It has never been used, in more than a century of the Permanent Court and the present Court, because States that submit a dispute to a court want the law applied and would otherwise settle.

3. "The International Court decided the continental shelf cases ex aequo et bono." Correct this statement.

It is wrong, and the error is the central one in the topic. In both the North Sea cases and Tunisia/Libya the Court decided according to law. The law of maritime delimitation itself requires that the result be equitable, so applying equitable principles is applying the law, not departing from it. In Tunisia/Libya the parties expressly asked the Court by special agreement to take account of equitable principles, the relevant circumstances and the new trends at the Third UN Conference, and expressly did not confer power to decide ex aequo et bono. Articles 74(1) and 83(1) of the Convention on the Law of the Sea now state the rule as delimitation by agreement on the basis of international law in order to achieve an equitable solution.

4. Does an equitable delimitation mean an equal one?

No. In the North Sea cases the Court rejected the doctrine of a just and equitable share, holding that its task was to delimit areas already appertaining to each State and not to apportion them de novo, because the rights of the coastal State in the shelf constituting the natural prolongation of its land territory exist ipso facto and ab initio by virtue of its sovereignty over the land. Proportionality between the areas allotted and the lengths of the relevant coasts is used, as in Tunisia/Libya, as a test of the equity of a result, not as a rule of entitlement.

5. Why is equity criticised as a basis for decision?

Because it is unpredictable and hard to police. A rule requiring an equitable result in the light of all the relevant circumstances gives a tribunal wide latitude and gives advisers little guidance, and because equity requires no proof of practice or opinio juris, a tribunal that has decided on other grounds can usually find circumstances to support it. The response has been to structure the discretion: courts now approach maritime delimitation through a defined sequence rather than an open weighing.

6. Is equity a source of international law?

Not under article 38(1), which lists conventions, custom, general principles and, as subsidiary means, judicial decisions and writings. Equity appears only in article 38(2), as a power to decide outside the law with the parties' consent. Its everyday operation is inside the other sources: it guides the application of rules, it fills gaps in a way that shades into the general principles of law, and in some fields, above all maritime delimitation and shared resources, the applicable rule itself demands an equitable result, so that applying equity is applying the law.

Article 18 governs voting: each Member has one vote, and decisions on important questions require a two-thirds majority of the members present and voting, the list of important questions including recommendations on international peace and security, elections, admission, suspension and expulsion of Members, questions relating to the trusteeship system and budgetary questions.

So the accurate statement is: binding on the Organization, recommendatory to the Members.

How a recommendation can still matter in law

This is the substance of the topic, and there are four routes.

As evidence of custom

Custom needs practice and opinio juris. A resolution can supply evidence of the second and sometimes of the first.

The vote of a State for a resolution is a statement by that State, in a formal setting, of its position on a question of law. When a resolution is adopted unanimously or nearly so, and when its language is normative rather than hortatory, and when States afterwards behave consistently with it, the resolution becomes strong evidence that a customary rule exists.

The test to apply is the one from the North Sea Continental Shelf cases, worked in [Custom]. Ask whether the text is of a norm-creating character, whether participation is widespread and representative including the States specially affected, and whether practice has been extensive and virtually uniform with the belief that law required it. A resolution adopted by acclamation and ignored in practice fails the third limb, however impressive the vote.

As a declaration of existing law

Some resolutions are drafted as statements of what the law already is, and are titled Declarations for that reason.

Resolution 2625 (XXV) of 24 October 1970, the Declaration on Principles of International Law concerning Friendly Relations and Co-operation among States in accordance with the Charter of the United Nations, is the outstanding example. It was adopted without a vote after years of negotiation, and it elaborates seven principles: the prohibition of the threat or use of force; the peaceful settlement of disputes; non-intervention in matters within the domestic jurisdiction of any State; the duty of States to co-operate; the equal rights and self-determination of peoples; the sovereign equality of States; and the duty to fulfil in good faith the obligations assumed in accordance with the Charter. It is universally treated as an authoritative statement of how the Charter's article 2 is to be read.

Resolution 217 A (III) of 1948, the Universal Declaration of Human Rights, is the other great example. It was not binding when adopted, and much of it is now accepted as customary.

Resolution 1514 (XV) of 1960, the Declaration on the Granting of Independence to Colonial Countries and Peoples, is expressly referred to in article 140 of the Convention on the Law of the Sea, which is a striking measure of how far a resolution can travel.

As a step in a treaty's life

Many conventions begin as resolutions. The Assembly adopts a declaration of principles, a negotiating conference follows, and a treaty results. Resolution 1962 (XVIII) of 1963, the Declaration of Legal Principles Governing the Activities of States in the Exploration and Use of Outer Space, was turned almost word for word into the Outer Space Treaty of 1967, which is [International Space Law].

As soft law

Some resolutions are meant to guide conduct without binding: codes of conduct, sets of principles, standards. They are not law and they are not nothing. States comply with them, invoke them against each other, and build them into national legislation, and over time some of them harden.

The developing-country argument, and why it failed

In the 1960s and 1970s newly independent States, then a large majority in the Assembly, argued that Assembly resolutions adopted by an overwhelming majority should be treated as making law. The argument had a real basis: those States had had no part in making the customary law they inherited, and the Assembly was the one place where each State had one vote.

The high point was the New International Economic Order resolutions of 1974, including the Charter of Economic Rights and Duties of States, which asserted, among other things, a right to nationalise foreign property with compensation determined by the nationalising State's own law.

It failed, and the reason is the second limb of the North Sea test. The developed States, whose interests were specially affected, voted against or abstained, and their practice did not follow. Without extensive and virtually uniform practice including the specially affected States, no custom formed, whatever the size of the majority.

The episode is worth knowing because it shows exactly what a resolution can and cannot do: it can express what a majority of States believe the law should be, and it cannot by itself make that belief into law.

A worked example

The General Assembly adopts, by 150 votes to 3 with 20 abstentions, a Declaration on the Protection of Undersea Cables, stating that every State is under an obligation to prosecute persons who damage a submarine cable. Two years later State T's court refuses to prosecute, and State U complains that international law has been broken.

Is the Declaration binding? No. Under articles 10 and 13 the Assembly may only recommend, and a Declaration is a recommendation whatever it is called.

Did State T vote for it? If it did, that is evidence of its own view of the law and may found an argument that it is precluded from denying the rule. If it voted against, that is evidence the other way, and if it has objected consistently since, it may be a persistent objector.

Has the Declaration made custom? Apply the North Sea test. The language, an obligation to prosecute, is norm-creating. A vote of 150 to 3 is widespread and representative, but the question is whether the major cable-laying and cable-owning States are among the 150 or among the 3 and the 20, because they are the specially affected States. And what have States actually done in the two years? If several have legislated and prosecuted, the practice limb is being satisfied; if nothing has happened anywhere, it is not.

What is the honest advice to State U? That the Declaration is powerful evidence and not a rule, that its case depends on practice since 1974-style majorities do not by themselves make law, and that the strongest single piece of material would be State T's own vote and any legislation State T has passed in reliance on it.

Distinctions table

What it does NOT mean

"Not binding" does not mean "not important". Resolution 2625 is not binding and is cited by the International Court as the authoritative gloss on article 2 of the Charter.

A large majority does not make law. The New International Economic Order resolutions show it. What matters is whether practice, including that of specially affected States, follows.

A resolution binding on the Organization is not binding on international law. A budget assessment binds a Member to pay; it creates no rule for anybody.

Not everything the Assembly adopts is a Declaration. Most resolutions are routine, and the ones that matter for this topic are the small number drafted in normative language and adopted with overwhelming support.

Limits and criticism

The category is unstable. There is no accepted test for when a resolution has crossed from recommendation into custom, so the same resolution can be described as declaratory by one State and aspirational by another, and both can point to respectable authority.

Voting is cheap. A State can vote for a resolution costlessly and behave otherwise, which is why practice, not votes, has to carry the weight. That is also why the second and third limbs of the North Sea test do the real work.

Article 38 has never been amended to accommodate any of this. A student is therefore asked to fit into a 1920 list a phenomenon that did not exist when the list was drafted, and the fit is imperfect by construction.

Quick revision

Articles 10 to 14 of the Charter: the Assembly may discuss, consider and recommend. Article 13 is the codification provision, under which the International Law Commission exists. Article 12 bars recommendations while the Council is seised, unless the Council requests.

Contrast article 25: Members agree to accept and carry out the decisions of the Security Council.

Binding on the Organization, recommendatory to Members. Article 17, budget and apportionment of expenses. Article 18: one vote each; two-thirds of members present and voting on important questions.

Four ways a resolution matters: evidence of custom, especially of opinio juris; a declaration of existing law; a step towards a treaty; and soft law.

Resolution 2625 (XXV), 24 October 1970, Friendly Relations Declaration, adopted without a vote: seven principles, including the prohibition of force, peaceful settlement, non-intervention, co-operation, self-determination, sovereign equality, and good faith.

The New International Economic Order resolutions failed to make law because the specially affected States voted against and their practice did not follow: the second and third limbs of the North Sea test.

Test yourself

1. Are resolutions of the General Assembly binding? Give the Charter basis.

Not on Members. Articles 10, 11, 13 and 14 empower the Assembly to discuss, consider and make recommendations, and article 12 restricts even that while the Security Council is seised of a dispute. The contrast is article 25, by which Members agree to accept and carry out decisions of the Security Council. Resolutions on the Organization's own internal affairs do bind: the budget and the apportionment of expenses under article 17, admissions, elections and rules of procedure.

2. How can a non-binding resolution acquire legal effect?

Four ways. It can be evidence of custom, particularly of opinio juris, since a State's vote is a formal statement of its view of the law. It can be a declaration of law that already exists, as Resolution 2625 is of the principles in article 2 of the Charter. It can be a step towards a treaty, as Resolution 1962 (XVIII) of 1963 was towards the Outer Space Treaty. And it can operate as soft law, guiding conduct and hardening over time.

3. What test decides whether a resolution has produced a customary rule?

The test from the North Sea Continental Shelf cases, applied to the resolution's text and its aftermath. The provision must be of a norm-creating character rather than hortatory. Support must be widespread and representative, including the States whose interests are specially affected. And State practice must have been extensive and virtually uniform in the sense of the rule, accompanied by a general recognition that a rule of law was involved. A resolution adopted by acclamation and ignored in practice satisfies the first two limbs and fails the third.

4. Why did the New International Economic Order resolutions fail to make law, despite overwhelming majorities?

Because a majority in the Assembly is not the test. The developed States, whose interests were specially affected by rules on nationalisation and compensation, voted against or abstained, and their subsequent practice did not conform. Without extensive and virtually uniform practice including the specially affected States, no customary rule formed. The episode is the clearest demonstration that resolutions express what States think the law should be and cannot by themselves make it so.

5. State the seven principles of Resolution 2625 (XXV).

The prohibition of the threat or use of force against the territorial integrity or political independence of any State; the settlement of international disputes by peaceful means; the duty not to intervene in matters within the domestic jurisdiction of any State; the duty of States to co-operate with one another; the principle of equal rights and self-determination of peoples; the sovereign equality of States; and the duty to fulfil in good faith the obligations assumed in accordance with the Charter.

6. Distinguish the legal effect of a General Assembly resolution from that of a Security Council resolution.

An Assembly resolution is a recommendation to Members under articles 10 to 14, and binds only in the Organization's internal affairs such as the budget under article 17. A Security Council resolution taken under Chapter VII is binding: article 25 provides that Members agree to accept and carry out the decisions of the Council in accordance with the Charter, and article 48 requires the action required to carry out those decisions to be taken by the Members. The difference is one of legal quality and not of importance: Resolution 2625 binds nobody and is treated as an authoritative statement of the Charter's principles.

It follows that a rule of international law is part of national law without any act of transformation, that national courts should apply it, and that in case of conflict international law prevails, because on this view it is the higher of the two.

Its two versions

Kelsen's version is the strict one. Every norm derives its validity from a higher norm, and the chain ends in a Grundnorm. Kelsen located the basic norm in international law, so municipal legal orders derive their validity from it. That is the position described in [Theories of the Basis of International Law].

The natural-law version holds that both systems derive from the same higher law, so neither can contradict the other.

What is right about it

It reflects how the international plane actually works. Article 27 means that from the outside a State's internal law is simply a fact and never a justification.

It also reflects how many States actually behave with custom: a great many legal systems, including the English and the Indian, apply customary international law without any legislation, which is a monist practice whatever the theory says.

What is wrong with it

No State applies it fully. Every legal system reserves the power of its legislature to override, and a monism that lets an unincorporated treaty defeat a statute would put the treaty-making executive above the legislature. That is a constitutional objection, not a technical one.

The two systems have different subjects. International law binds States primarily; municipal law binds individuals. Treating them as one order requires the claim that international law ultimately regulates individuals too, which is arguable but not obvious.

It cannot explain the transformation practice. If international law were automatically part of national law, the elaborate machinery States maintain for incorporating treaties would be pointless.

Dualism

What it says

The two are separate legal systems, and the classic statement of the differences comes from Triepel and Anzilotti.

Different sources. International law comes from the common will of States; municipal law comes from the will of a single State.

Different subjects. International law regulates States; municipal law regulates individuals.

Different subject matter. International law governs relations between States; municipal law governs relations within a State.

It follows that a rule of international law can have no effect inside a State until that State makes it part of its own law, by legislation or otherwise. That process is transformation.

What is right about it

It matches most States' actual constitutional practice for treaties. In India, in the United Kingdom, in Canada, in Australia, a treaty concluded by the executive does not by itself change the law that courts apply.

It also protects the legislature. Since the treaty-making power is executive almost everywhere, unmediated monism would let the executive legislate by treaty.

What is wrong with it

It is too tidy about subjects. Individuals now have rights and duties directly under international law, as [Subjects of International Law] shows, so the clean division between systems by whom they bind is no longer accurate.

It does not describe how custom is treated. Most dualist States apply customary international law without transformation.

It can produce embarrassment. A State can be in clear breach on the international plane while its own courts are obliged to apply the statute that produced the breach.

The doctrines that do the work

The theories are labels; these three doctrines are what a court actually applies.

The doctrine of incorporation, sometimes called adoption. Rules of international law are part of the law of the land automatically, unless they conflict with an Act of Parliament or with settled judicial decision. This is the general position for custom in India and in England.

The doctrine of transformation. Rules of international law are not part of the law of the land unless they have been made so by legislation, judicial decision or long established usage. This is the general position for treaties in India and in England.

Consistent interpretation, sometimes called the presumption of conformity. Where a statute is capable of two readings, a court prefers the reading that conforms to the State's international obligations, because the legislature is presumed not to intend a breach. This is the most practically useful of the three, because it operates even where a treaty has not been incorporated, and its Indian statement is in [International Law in Indian Courts].

Comparative practice, briefly

The United Kingdom. Custom is incorporated; treaties require legislation; a statute prevails over both.

The United States. Article VI of the Constitution makes treaties made under the authority of the United States part of the supreme law of the land, which looks strongly monist. The effect is cut down by the distinction between self-executing and non-self-executing treaties, the latter needing implementing legislation, and by the rule that a later statute prevails over an earlier treaty as a matter of domestic law.

The Netherlands and several civil-law States. Constitutions provide expressly that treaties, once published, have effect and may prevail even over the constitution in defined cases. This is the closest practice comes to genuine monism.

India is dealt with in the next chapter.

A worked example

State D, a dualist State, ratifies a convention requiring it to make torture a criminal offence. Parliament passes no legislation. Two years later an official of State D tortures a detainee. The detainee sues in State D's courts, relying on the convention.

On the international plane. State D is in breach. Article 26 requires performance in good faith and article 27 prevents it pleading the absence of legislation as an excuse. Another State party may invoke its responsibility.

In State D's courts. The convention is not part of the law of the land, so the court cannot convict on it or award damages under it. That is dualism doing what it does.

Is that the end? No, and this is where a good answer separates itself. If any existing statute or constitutional guarantee is capable of a reading that conforms to the convention, the court should prefer that reading. If State D's constitution guarantees life and personal liberty, the convention can inform its content without being incorporated. And if the prohibition of torture is a rule of customary international law, then in a State that incorporates custom the court may apply it directly, since the barrier is to unincorporated treaties and not to custom.

Change the facts: State D's own statute expressly authorises the conduct. Now consistent interpretation is unavailable, because the statute is not capable of two readings. The court must apply the statute, and State D remains in breach internationally. That is the embarrassment dualism produces, and the remedy is political rather than judicial.

Distinctions table

What it does NOT mean

No State is purely monist or purely dualist. Almost every system incorporates custom and transforms treaties, which is a mixture. Describing India or England as "dualist" without that qualification is a half-truth.

Dualism does not mean international law is ignored. It means it operates on a different plane, and it reaches national law through incorporation, through legislation, and above all through consistent interpretation.

A treaty that has not been incorporated is not useless in a national court. It can be used to construe a statute, to inform a constitutional guarantee, and to resolve ambiguity.

Monism does not mean a treaty always beats a statute. Even in strongly monist systems the effect depends on constitutional provisions about self-executing treaties and about which prevails on conflict.

Limits and criticism

The debate is largely about description, not decision. Courts do not decide cases by choosing a theory; they apply their own constitution's rules about treaties and custom, and the theories are labels attached afterwards.

The labels obscure the interesting question. What matters is not whether a system is monist but which rules of international law are directly applicable, at whose instance, and with what rank. Those questions have different answers for custom, for treaties, for peremptory norms and for decisions of international organisations, and no single label captures the pattern.

The gap the debate leaves is real. In dualist practice a State can be in breach internationally while its courts are compelled to apply the offending statute, and neither theory solves that; only legislation does.

Quick revision

Monism: one legal order; international law is automatically part of national law and prevails on conflict. Kelsen, with the Grundnorm in international law.

Dualism: two systems, differing in sources, subjects and subject matter; international law applies internally only after transformation. Triepel and Anzilotti.

Article 27 VCLT: a party may not invoke its internal law as justification for failure to perform a treaty. Article 46: consent given in violation of internal law on competence may be invoked only if the violation was manifest and concerned a rule of fundamental importance.

Three working doctrines: incorporation (custom, automatically), transformation (treaties, by legislation), and consistent interpretation, which operates even on an unincorporated treaty.

Practice: the United Kingdom incorporates custom and transforms treaties; the United States makes treaties supreme law but distinguishes self-executing from non-self-executing; several civil-law constitutions give treaties direct effect.

No State is purely one or the other.

Test yourself

1. State the monist and dualist positions and name their chief exponents.

Monism holds that international law and municipal law form a single legal order, so that a rule of international law is part of national law without transformation and prevails in case of conflict; Kelsen is its chief exponent, locating the Grundnorm of the whole legal order in international law. Dualism holds that they are separate systems differing in their sources, their subjects and their subject matter, so that international law operates within a State only when transformed into national law; Triepel and Anzilotti are its chief exponents.

2. Can a State plead its own constitution as a defence to a breach of treaty?

No. Article 27 of the Vienna Convention provides that a party may not invoke the provisions of its internal law as justification for its failure to perform a treaty. The only qualification is article 46, which looks the other way: a State may not invoke a violation of its internal law on competence to conclude treaties as invalidating its consent unless the violation was manifest and concerned a rule of internal law of fundamental importance, a violation being manifest if it would be objectively evident to any State conducting itself in accordance with normal practice and in good faith.

3. Distinguish incorporation from transformation.

Under incorporation, rules of international law become part of the law of the land automatically and are applied by national courts unless they conflict with a statute or settled judicial decision. Under transformation, they are not part of the law of the land unless made so by legislation, judicial decision or long established usage. Most systems, including the Indian and the English, incorporate custom and transform treaties, so the two doctrines operate side by side within one legal system on different classes of rule.

4. A treaty has been ratified but not implemented by legislation. What use is it in a national court of a dualist State?

It cannot be the direct source of a right or an offence, because it is not part of the law of the land. It remains useful in three ways. It can be used to construe an ambiguous statute, on the presumption that the legislature does not intend to legislate contrary to the State's international obligations. It can inform the content of a constitutional guarantee expressed in general terms. And where its content is also a rule of customary international law, the court may apply that rule directly if the system incorporates custom, because the barrier is to unincorporated treaties and not to custom.

5. "Neither monism nor dualism describes any real legal system." Discuss.

The statement is close to true. Almost every system applies customary international law without legislation, which is monist, and requires legislation before a treaty changes the law courts apply, which is dualist. The United States is often called monist because article VI makes treaties the supreme law of the land, yet it distinguishes self-executing from non-self-executing treaties and lets a later statute prevail domestically over an earlier treaty. The theories are best treated as descriptions of two tendencies, and the useful questions are which rules are directly applicable, at whose instance, and with what rank, which have different answers for custom, treaties and peremptory norms.

6. What is the presumption of conformity, and why is it the most useful of the three doctrines?

It is the rule of construction that where a statute is capable of more than one meaning, a court prefers the meaning that conforms to the State's international obligations, because the legislature is presumed not to intend a breach. It is the most useful because it operates without incorporation and without legislation, so it reaches unincorporated treaties, and because it applies in the great majority of cases, where the question is what a statute means rather than whether it is valid. Its limit is that it cannot be used where the statute admits of only one reading, and in that case the court must apply the statute and the State remains in breach.

Putting article 73 and entry 14 together: the Union executive may conclude a treaty without Parliament. No ratification by either House is constitutionally required, and there is no Indian equivalent of the United States Senate's advice and consent.

That is a large power, and the check on it is the next section: concluding a treaty and making it enforceable are two different things.

Who makes treaties enforceable: Parliament

Article 253 provides that notwithstanding anything in the foregoing provisions of Chapter I of Part XI, Parliament has power to make any law for the whole or any part of the territory of India for implementing any treaty, agreement or convention with any other country or countries or any decision made at any international conference, association or other body.

Two consequences, and the second is the one that surprises students.

Implementation is a matter for Parliament, not for the executive that made the treaty. So a treaty that requires a change in Indian law, that creates an offence, or that affects private rights needs an Act.

Article 253 overrides the federal division. The words "notwithstanding anything in the foregoing provisions of this Chapter" mean Parliament may legislate to implement a treaty even on a subject in the State List. That is a significant expansion of Union power, and it is the constitutional route by which a wide range of international obligations, on the environment, on human rights, on trade, have entered Indian law.

Familiar examples of implementing legislation include the Territorial Waters, Continental Shelf, Exclusive Economic Zone and Other Maritime Zones Act 1976 for the law of the sea, dealt with in [The Law of the Sea], and the Arbitration and Conciliation Act 1996 for the UNCITRAL Model Law and the New York Convention, dealt with in [UNCITRAL and International Commercial Arbitration] and [Enforcement of Foreign Arbitral Awards].

What an Indian court actually does

With custom: incorporation

Customary international law is applied by Indian courts without any legislation, provided it does not conflict with an Act of Parliament. The authority is the case below, and this is the doctrine of incorporation described in [International Law and Municipal Law].

With treaties: transformation

A treaty that India has ratified but Parliament has not implemented does not by itself alter the law that an Indian court applies. It cannot create an offence, and it cannot by itself create or take away a private right.

With ambiguity: consistent interpretation

Where a statute or a constitutional guarantee is capable of more than one reading, an Indian court prefers the reading that conforms to India's international obligations. This is the rule that does most of the work in practice, and it is stated in the case below as a rule of construction: courts must interpret national law in a way that avoids confrontation with international law.

The leading case

Facts. Gramophone Company of India Ltd v. Birendra Bahadur Pandey, AIR 1984 SC 667, also reported at (1984) 2 SCC 534 and 1984 SCR (2) 664, decided on 21 February 1984 by Chinnappa Reddy, Venkataramiah and Misra JJ. A consignment of pre-recorded cassettes was shipped from Singapore to Kathmandu in Nepal and arrived at Calcutta in transit. The Gramophone Company, alleging that the cassettes were pirated copies infringing its copyright, invoked section 53 of the Copyright Act 1957. The question was whether goods merely passing through India in transit to a land-locked State were imported within sections 51 and 53 of the Act. That turned on whether Nepal's right of innocent passage of goods across Indian territory, recognised in international law and in treaty, cut down the ordinary meaning of the word.

Held. Allowing the appeal, the Court held the goods were imported for the purposes of the Act. On the general question it stated the Indian position. There are two schools of thought: under the doctrine of incorporation, rules of international law are incorporated into the law of the land automatically and are considered part of it unless in conflict with an Act of Parliament; under the doctrine of transformation they are not part of the law of the land unless already made so by an Act of Parliament, a judicial decision or long established usage. The Court held that nations must march with the international community and that municipal law must respect rules of international law even as nations respect international opinion. The comity of nations requires that rules of international law be accommodated in municipal law even without express legislative sanction, provided they do not run into conflict with Acts of Parliament. But where they do run into such conflict, the sovereignty and integrity of the Republic and the supremacy of the constituted legislatures in making the laws are not to be subjected to external rules. The Court added the rule of construction that courts must interpret national law, wherever possible, so as to avoid confrontation with international law.

Why it matters. It is the statement an Indian answer to MU's topic 1.6 has to contain, and it contains three propositions, not one: custom is accommodated without legislation; an Act of Parliament prevails where they conflict; and national law is to be construed, so far as it will bear it, to avoid confrontation.

A worked example

India ratifies the Convention on the Protection of Coastal Wetlands, which requires each party to prohibit construction within five hundred metres of a designated wetland. Parliament passes no Act. A developer begins building at four hundred metres from a designated wetland, with a permit issued under a State law. A citizens' group sues.

Can the group rely on the Convention directly? No. India follows transformation for treaties, so an unimplemented convention does not by itself create an enforceable prohibition, and a court cannot make lawful conduct unlawful by reference to it.

Is the Convention therefore irrelevant? No, and this is the substance of the answer. If the State law or the permitting rules are capable of more than one reading, the court should prefer the reading that conforms to the Convention, on the Gramophone rule of construction. If a constitutional guarantee is engaged, the Convention can inform its content without being incorporated.

Could Parliament legislate even though land and local government are State subjects? Yes. That is precisely what article 253 is for: Parliament may make any law for implementing a treaty notwithstanding the distribution of legislative powers.

Suppose instead the obligation were a rule of customary international law. Then the court could apply it directly on the doctrine of incorporation, subject to the limit Gramophone states: it must not conflict with an Act of Parliament.

Suppose an Act of Parliament expressly permits construction at four hundred metres. Then neither route is open. The Act prevails in an Indian court, and India is in breach on the international plane, where article 27 of the Vienna Convention prevents it pleading its own legislation as an excuse. The remedy is legislative, not judicial.

Distinctions table

What it does NOT mean

Article 51 is not enforceable. It is a Directive Principle, and article 37 makes the principles unenforceable by any court while declaring them fundamental in the governance of the country.

Treaty-making does not require Parliament. It requires the Union executive, under article 73 with entry 14. Implementation requires Parliament, under article 253. Confusing the two is the commonest error in this topic.

Article 253 does not make treaties part of Indian law. It gives Parliament a power. Until the power is exercised, nothing has changed in Indian law.

"Dualist" overstates it. India transforms treaties and incorporates custom, and its courts construe statutes to conform. The accurate description is a modified dualism, not a pure one.

Limits and criticism

Parliament has no formal role in treaty-making. The executive can bind India internationally on any subject, including subjects in the State List, and Parliament's involvement comes only at the implementation stage, if at all. Proposals for a statutory requirement of parliamentary scrutiny have been made repeatedly and none has been enacted.

Article 253 unsettles the federal balance. A treaty concluded by the Union executive can open the door to Union legislation on a State subject, and the States have no constitutional voice in whether the treaty is concluded.

The gap between ratification and implementation is real. India ratifies conventions that are not implemented for years or at all, and in that period the obligation binds India internationally while being unenforceable in its own courts.

The presumption of conformity has a hard limit. It works only where the statute will bear the conforming reading. Where the words admit of one meaning only, the court must apply them, India remains in breach, and the position cannot be cured by construction.

Quick revision

Article 51: the State shall endeavour to promote international peace and security, maintain just and honourable relations between nations, foster respect for international law and treaty obligations, and encourage settlement of disputes by arbitration. A Directive Principle, unenforceable under article 37, and used all the same.

Treaty-making is executive: article 73 plus List I entries 10 to 14, entry 14 being entering into and implementing treaties. No parliamentary approval is needed.

Implementation is parliamentary: article 253, which lets Parliament legislate to implement a treaty notwithstanding the distribution of legislative powers, so even on a State subject.

Custom is applied by incorporation, treaties by transformation, and ambiguity is resolved by consistent interpretation.

Gramophone Company of India Ltd v. Birendra Bahadur Pandey, AIR 1984 SC 667, (1984) 2 SCC 534, 21 February 1984: the comity of nations requires rules of international law to be accommodated in municipal law even without express legislative sanction, provided they do not conflict with Acts of Parliament; where they conflict, the supremacy of the legislature prevails; and courts must interpret national law so as to avoid confrontation with international law.

Test yourself

1. Who can make a treaty on behalf of India, and does Parliament have to approve it?

The Union executive, under article 73 of the Constitution read with entry 14 of List I of the Seventh Schedule, which covers entering into treaties and agreements with foreign countries. Parliament has no constitutional role in the making of a treaty and its approval is not required. Parliament's power under article 253 is to legislate for implementing a treaty, which is a different stage and may never be reached.

2. Explain article 253 and its effect on the federal structure.

Article 253 provides that notwithstanding anything in the foregoing provisions of that Chapter, Parliament has power to make any law for the whole or any part of India for implementing any treaty, agreement or convention with any other country, or any decision made at an international conference, association or other body. Its effect is that Parliament may legislate even on a subject in the State List where it is implementing a treaty. Since the treaty is concluded by the Union executive without the States' concurrence, the provision allows the Union to acquire legislative competence over State subjects by the act of making a treaty, which is a significant qualification of the federal division.

3. State the three propositions in Gramophone.

First, that the comity of nations requires rules of international law to be accommodated in municipal law even without express legislative sanction, provided they do not run into conflict with Acts of Parliament, which is the doctrine of incorporation applied to custom. Second, that where they do conflict, the sovereignty and integrity of the Republic and the supremacy of the constituted legislatures in making the laws are not subject to external rules, so the Act prevails. Third, that courts must interpret national law, wherever possible, so as to avoid confrontation with international law.

4. Is article 51 of any use if it cannot be enforced?

Yes. Article 37 makes the Directive Principles unenforceable by any court while declaring them fundamental in the governance of the country and a duty of the State to apply in making laws. Article 51 is accordingly used as an interpretive direction: it supports reading statutes and constitutional guarantees, where they are capable of it, so as to conform to India's international obligations, and it is the constitutional foundation of the rule of construction stated in Gramophone.

5. India ratifies a convention creating a new criminal offence. No Act is passed. Can a person be prosecuted under the convention in India?

No. India follows transformation for treaties, so an unimplemented convention is not part of the law of the land and cannot create an offence. Article 20(1) of the Constitution independently forbids conviction except for the violation of a law in force at the time of the act. Prosecution requires an Act of Parliament, which article 253 empowers Parliament to pass. Until then India is in breach of the convention on the international plane, and article 27 of the Vienna Convention prevents it pleading the absence of legislation as a justification.

6. Would you describe India as monist or dualist? Justify your answer.

Neither label fits without qualification, and the accurate description is a modified dualism. India transforms treaties: an unimplemented convention does not change the law that Indian courts apply, and article 253 exists precisely because implementation needs an Act. India incorporates custom: on the authority of Gramophone, rules of international law are accommodated in municipal law without express legislative sanction so long as they do not conflict with an Act of Parliament. And Indian courts apply the presumption of conformity, construing national law to avoid confrontation with international law. That mixture, dualist for treaties and monist for custom with an interpretive bridge between them, is the practice of most common law systems.

The two theories

This is the heart of MU's topic and an essay question in its own right.

The constitutive theory

Recognition creates the State. An entity becomes a subject of international law, with rights and duties, only when existing States recognise it. Until then it is a fact and not a person in law.

The argument for it. International law rests on consent, and a new State cannot acquire rights against existing States without their acceptance. There is also a practical point: somebody has to decide, and there is no international authority that certifies statehood, so the decision falls to States individually.

The objections, and they are strong.

It produces relativity. If recognition constitutes, an entity is a State as against those that have recognised it and not as against those that have not. Israel would be a State in Delhi and not a State in some other capital, which is an unworkable account of legal personality.

It leaves the unrecognised entity outside the law. If an entity is not a State until recognised, it owes no obligations. It could not commit aggression, because aggression is a wrong of States. Nobody accepts that conclusion.

Recognition is a political act. States recognise for reasons of policy. Making legal existence turn on political choice puts the law at the disposal of politics.

How many recognitions? The theory gives no answer. One? A majority? Which States count?

The declaratory theory

Recognition acknowledges an existing fact. An entity that satisfies the criteria is a State, and recognition merely declares what is already the case and opens diplomatic relations.

The argument for it. It matches the Convention itself: article 3 of Montevideo provides that the political existence of the State is independent of recognition by the other States. It avoids relativity. And it fits practice: unrecognised entities are treated as bound by international law, and States that do not recognise one another still complain of each other's breaches in legal terms.

The objections. It understates what recognition actually does. An unrecognised State cannot make treaties with the States that refuse it, cannot maintain embassies with them, and its officials and property get no immunity in their courts. Saying its statehood is unaffected is true and thin.

The position to write

The declaratory theory is the better view and is the one modern practice supports, with two concessions that a good answer makes.

Recognition is strong evidence. Where the facts are doubtful, and they often are, widespread recognition is the best available proof that the criteria are satisfied, and refusal by most States is the best available proof that they are not.

Recognition has large practical consequences, set out in [Modes and Legal Effects of Recognition], even if it has no constitutive legal effect.

Membership of the United Nations is a different question

Students routinely confuse admission with recognition, and the distinction is easy marks.

Article 4 provides that membership is open to all peace-loving States which accept the obligations of the Charter and, in the judgment of the Organization, are able and willing to carry them out, and that admission is effected by a decision of the General Assembly upon the recommendation of the Security Council.

Three consequences.

Admission is not collective recognition. A Member does not thereby recognise every other Member, and States that do not recognise each other sit in the same Assembly.

The Security Council's recommendation means the veto applies. Admission is therefore a political process, and entities have been kept out for many years by a single negative vote.

A State that is not a Member is still a State. Switzerland was a State for centuries and joined only in 2002.

The duty of non-recognition

A State is generally free to recognise or not. There is one important qualification.

Where a situation has been created by a serious breach of a peremptory norm, in particular by the unlawful use of force or in violation of the right of self-determination, States are under a duty not to recognise it as lawful. Resolution 2625 states that the territory of a State shall not be the object of acquisition by another State resulting from the threat or use of force, and that no territorial acquisition resulting from the threat or use of force shall be recognised as legal, and the International Law Commission's Articles on State Responsibility carry the same obligation for serious breaches of peremptory norms, dealt with in [Circumstances Precluding Wrongfulness, and Reparation].

The classical formulation is the Stimson doctrine of 1932, by which the United States declined to recognise the situation created by Japan's occupation of Manchuria.

Recognition of belligerency and insurgency

Two lesser forms belong here because examiners ask about them alongside recognition of States.

Recognition of insurgency acknowledges that a revolt has grown beyond ordinary criminality and that the insurgents are to be treated as having a legal status for limited purposes. It does not confer belligerent rights.

Recognition of belligerency is more far-reaching and traditionally requires four conditions: a general armed conflict within the State, occupation and a measure of orderly administration of a substantial part of the territory by the insurgents, observance of the laws of war by forces under a responsible command, and a practical necessity for outside States to define their attitude. Its effect is that the law of armed conflict applies between the parties, the recognising State assumes the duties of neutrality, and the parent State ceases to be responsible for acts in the area the insurgents control.

A worked example

The province of Zenda declares independence from State Y. It has four million inhabitants, controls the whole province, has a functioning administration and police, and asks for recognition. State Y calls the declaration a rebellion. Twenty States recognise Zenda; ninety do not. Zenda applies to join the United Nations and State Y's ally on the Security Council votes against.

Is Zenda a State? Apply Montevideo. A permanent population: yes, four million. A defined territory: yes, the province, and the fact that State Y disputes the boundary does not matter, since a defined territory does not require settled frontiers. A government: yes, on the facts. Capacity to enter into relations: yes, if it is genuinely independent of State Y and is not being run from outside.

Does it matter that only twenty States have recognised it? On the declaratory theory, no: statehood follows the facts. On the constitutive theory, Zenda would be a State only as against those twenty, which is the relativity objection.

Does the pattern of recognition matter at all? Yes, as evidence. If ninety States withhold recognition because they doubt Zenda's independence, that is powerful evidence on the fourth criterion.

Does the failure to join the United Nations decide it? No. Admission is a separate question under article 4, requiring a recommendation of the Security Council and therefore exposed to the veto, and non-membership does not deprive an entity of statehood.

Change one fact: suppose Zenda was created by the army of a neighbouring State, which invaded and installed the administration. Now the fourth criterion fails, because Zenda is not independent, and the duty of non-recognition applies, since the situation results from the unlawful use of force. States are obliged not to recognise it as lawful, whatever its apparent effectiveness.

Distinctions table

What it does NOT mean

Recognition is not admission to the United Nations. Article 4 admission requires a Security Council recommendation and an Assembly decision; it is not collective recognition and it does not create statehood.

A defined territory does not mean settled boundaries. A State can exist with every frontier disputed.

Non-recognition does not mean the entity is outside the law. It remains bound by international law, and the duty of non-recognition of an unlawful situation is itself a legal duty.

Recognition is not the same as diplomatic relations. States can recognise one another and not exchange ambassadors, and breaking off diplomatic relations does not withdraw recognition.

Limits and criticism

The criteria are indeterminate at the margin. How much control is enough for a government, and how independent is independent, are questions the Montevideo formula does not answer, and they are exactly the questions that arise in every contested case.

The theories do not decide cases. In practice States recognise for reasons of policy and afterwards find the criteria satisfied, and the choice between constitutive and declaratory rarely changes an outcome.

Effectiveness and legality pull apart. The classical criteria are about effectiveness; the duty of non-recognition is about legality. Where an entity is effective but unlawfully created, the two point in opposite directions and the law has no tidy answer.

Quick revision

Montevideo Convention 1933, article 1: a State should possess a permanent population, a defined territory, government, and capacity to enter into relations with other States. A fifth requirement, lawful creation, is now generally added.

A defined territory does not require settled frontiers. Capacity to enter into relations means independence, and it is the criterion that distinguishes a State from a province.

Constitutive theory: recognition creates the State. Objections: relativity, an entity outside the law, politics deciding legal existence, and no answer to how many recognitions are needed.

Declaratory theory: recognition acknowledges a fact. Supported by article 3 of Montevideo: the political existence of the State is independent of recognition by the other States. The better view; recognition remains strong evidence and has large practical effects.

Article 4 of the Charter: admission of peace-loving States able and willing to carry out the obligations, by decision of the General Assembly upon the recommendation of the Security Council. Admission is not recognition, and non-membership is not a bar to statehood.

Duty of non-recognition where a situation results from a serious breach of a peremptory norm, in particular the unlawful use of force: Resolution 2625; historically the Stimson doctrine, 1932.

Recognition of belligerency: general armed conflict, occupation and orderly administration of substantial territory, forces under responsible command observing the laws of war, and a practical necessity for outside States to define their attitude.

Test yourself

1. State the criteria of statehood and explain which is the most important.

Article 1 of the Montevideo Convention 1933 requires a permanent population, a defined territory, government, and capacity to enter into relations with other States. The last is the most important, because it is what distinguishes a State from a component unit of a federation or a dependent territory: an entity that takes its orders from another State fails it however large its population or effective its administration. A fifth requirement, that the entity was lawfully created, is now generally added, since an entity established by unlawful force will not be recognised.

2. Distinguish the constitutive and declaratory theories, and say which is preferable.

The constitutive theory holds that recognition creates statehood, so an unrecognised entity is not a State in law. The declaratory theory holds that recognition merely acknowledges a state of affairs that already exists on the facts. The declaratory theory is preferable. The constitutive theory produces relativity, making an entity a State against some States and not others; it puts an unrecognised entity outside international law altogether, so that it could not even commit aggression; it makes legal existence turn on political choice; and it cannot say how many recognitions are required. The concession to be made is that recognition is powerful evidence where the facts are doubtful, and that it carries large practical consequences.

3. Does admission to the United Nations amount to recognition?

No. Article 4 of the Charter provides that membership is open to peace-loving States that accept the Charter obligations and are able and willing to carry them out, and that admission is by decision of the General Assembly upon the recommendation of the Security Council. A Member does not thereby recognise every other Member, and States that do not recognise one another sit together in the Assembly. Because a Council recommendation is required, admission is subject to the veto and is a political process, and an entity may be a State for many years without being a Member.

4. Can a State exist without settled boundaries?

Yes. The Montevideo criterion is a defined territory, not a delimited one, and what is required is a core of territory over which the entity exercises control. Many States have come into existence with frontier disputes outstanding and have been recognised and admitted to the United Nations while those disputes continued.

5. What is the duty of non-recognition?

The obligation not to recognise as lawful a situation created by a serious breach of a peremptory norm, in particular by the threat or use of force or in violation of self-determination. Resolution 2625 (XXV) provides that the territory of a State shall not be the object of acquisition by another State resulting from the threat or use of force, and that no territorial acquisition so resulting shall be recognised as legal, and the International Law Commission's Articles on State Responsibility impose a corresponding duty on all States in the case of serious breaches of peremptory norms. Its historical ancestor is the Stimson doctrine of 1932, by which the United States refused to recognise the situation Japan created in Manchuria.

6. What conditions must be satisfied before belligerency may be recognised, and what follows?

There must be a general armed conflict within the State; the insurgents must occupy and administer in an orderly way a substantial part of its territory; their forces must act under a responsible command and observe the laws of war; and there must be a practical necessity for outside States to define their attitude. Recognition then has three effects: the law of armed conflict applies between the parties, the recognising State assumes the duties of a neutral, and the parent State ceases to be internationally responsible for what happens in the area the insurgents control. Recognition of insurgency is a lesser step, acknowledging that the revolt is more than ordinary criminality without conferring belligerent rights.

The Estrada doctrine: no pronouncement at all

Stated by Genaro Estrada, the Foreign Minister of Mexico, in 1930. Its position is that the practice of granting or withholding recognition of governments is itself objectionable, because it amounts to passing judgment on the internal affairs of another State. Mexico therefore announced that it would not issue declarations of recognition of governments, and would confine itself to maintaining or withdrawing its diplomatic agents as it saw fit.

The idea behind it. Recognition of governments is an insult in either direction: to grant it is to claim a right to approve, and to withhold it is to intervene.

Its influence. This is now the dominant practice. Many States, including the United Kingdom since 1980, have announced that they no longer recognise governments as such, and instead decide whether to have dealings with a new authority and leave the inference to be drawn from the dealings themselves.

The objection. It does not remove the problem, it relocates it. Courts still have to decide whether an authority is the government of a State for the purposes of immunity, of the validity of its acts, and of who may sue on the State's behalf, and if the foreign ministry will not say, the court must work it out from the conduct of relations.

Effective control: the working test

Underneath both doctrines is the test that practice actually applies. A government is recognised, or dealt with, when it exercises effective control over the territory and population with a reasonable prospect of permanence, and is able and willing to carry out the State's international obligations.

That test is the analogue of the government criterion in [Recognition of States], and it is the one a candidate should give as the general rule, with Tobar and Estrada as the two positions that qualify it in opposite directions.

De facto and de jure recognition of governments

The distinction is set out generally in [Modes and Legal Effects of Recognition], and it belongs to governments rather than to States, so the outline is here.

De facto recognition acknowledges that an authority is in effective control, while reserving judgment on whether its position is lawful or permanent. It is provisional, it can be withdrawn, and it is what States use when a new regime looks likely to last but its future is uncertain.

De jure recognition acknowledges that the authority is the lawful government. It is intended to be permanent, and withdrawing it is a serious step.

Two governments of the same State may be recognised at once in the two capacities, one de jure and the other de facto, which happened repeatedly during and after civil wars in the twentieth century, and could not happen with States.

A worked example

The army of State K overthrows the elected government on 1 March. The President flees abroad and continues to claim to be the lawful head of State. By 1 June the army controls the whole country, the administration functions, and taxes are collected. State L must decide what to do.

Does State K still exist? Yes, and it is still recognised. A change of government does not touch the State, so State K's treaties with State L continue, its debts remain owed, and its responsibility for earlier wrongs is unaffected.

On the Tobar doctrine, State L should withhold recognition until the new authority is legitimated by a constitutional process, typically an election. The consequence is that State L has no one to deal with in a country where its nationals live and its investments sit.

On the Estrada doctrine, State L makes no pronouncement at all. It decides whether to keep its ambassador in place and leaves the inference to be drawn. That is the modern practice.

On effective control, the new authority satisfies the test by 1 June: it controls the territory, the administration works, and permanence is reasonably likely.

What if State L needs an answer for its own courts? Suppose the deposed President sues in State L's courts to recover State K's bank account. The court must decide who is entitled to act for State K. If State L has stopped making declarations of recognition, the court will look at whether State L's government deals with the new authority, whether it has kept its embassy, and whether it has accepted the authority's diplomats. That is exactly the difficulty the Estrada doctrine creates: the political question does not disappear, it lands on the judge.

Distinctions table

What it does NOT mean

Recognising a government is not approving it. States have always said so, and the Estrada doctrine exists because saying so was not believed.

Non-recognition does not suspend the State's obligations. The State remains bound whether or not anybody recognises the authority that governs it.

Effective control is not permanent control. The test asks for a reasonable prospect of permanence, not a guarantee.

A government in exile is not a government by effective control. It may be recognised for political reasons, and the recognition is a departure from the ordinary test rather than an application of it.

Limits and criticism

Recognition of governments has always been political and has never stopped being so. The Estrada doctrine is a confession that the legal test could not be kept free of policy, and abolishing the declaration does not abolish the judgment.

The costs of non-recognition fall on private people. Contracts, marriages, births and judgments in a territory governed by an unrecognised authority have to be dealt with by courts elsewhere, and refusing to acknowledge the authority means refusing to acknowledge the ordinary transactions of a whole population.

The two doctrines answer different questions. Tobar asks whether a government ought to be dealt with; Estrada asks whether that question is anyone else's business. Presented as rivals on the same question, as they usually are in textbooks, both look weaker than they are.

Quick revision

Recognition of a State is about existence and happens once. Recognition of a government is about who represents it and arises whenever power is taken outside the constitutional process.

A change of government does not change the State: treaties, debts and responsibility continue, because the State is the same legal person.

Tobar doctrine (Ecuador, 1907): no recognition of a government coming to power by revolution until legitimated by constitutional process. The Wilson doctrine is its American counterpart. Objection: it is intervention, and it fails against effective power.

Estrada doctrine (Mexico, 1930): make no pronouncement at all; maintain or withdraw diplomatic agents and let the inference be drawn. Now the dominant practice, followed by the United Kingdom since 1980. Objection: the question moves from the foreign ministry to the courts.

Effective control is the working test: control of territory and population, a reasonable prospect of permanence, and ability and willingness to carry out international obligations.

De facto recognition is provisional and acknowledges control; de jure acknowledges lawfulness and is intended to be permanent. Two governments of one State may be recognised at once in the two capacities.

Test yourself

1. Distinguish recognition of a State from recognition of a government.

Recognition of a State acknowledges that an entity exists as a State; it happens once and is not repeated. Recognition of a government acknowledges that a particular authority is competent to represent a State whose existence is already accepted, and the question arises again whenever a government takes power outside the constitutional process. Recognising a State implies recognising its government, but not the reverse, and withdrawing recognition from a government leaves the State's recognition untouched.

2. Does a change of government affect a State's treaties and debts?

No. The State is the same legal person before and after, so its treaties continue to bind it, its debts remain owed, and its responsibility for earlier internationally wrongful acts is unaffected. That is the principle of the continuity of States, and it is what makes recognition of governments a question about representation rather than about obligation.

3. State the Tobar and Estrada doctrines and say which reflects modern practice.

The Tobar doctrine, proposed by Ecuador's Foreign Minister in 1907, is that a government coming to power by revolution or coup should not be recognised until legitimated by its own country's constitutional process; the Wilson doctrine applied a similar test from 1913. The Estrada doctrine, stated by Mexico's Foreign Minister in 1930, is that granting or withholding recognition of governments is itself objectionable as a judgment on another State's internal affairs, so no declaration should be made and the State should confine itself to maintaining or withdrawing its diplomatic agents. Estrada reflects modern practice: many States, including the United Kingdom since 1980, no longer recognise governments as such.

4. What is the working test for recognising a government?

Effective control: the authority must control the territory and population, with a reasonable prospect of permanence, and must be able and willing to carry out the State's international obligations. It is the same enquiry as the government criterion for statehood, applied to a State that already exists.

5. What difficulty does the Estrada doctrine create for national courts?

It removes the answer without removing the question. A court still has to decide who represents a foreign State for the purposes of sovereign immunity, of the validity of the State's acts, and of who may sue and be sued on its behalf. If the foreign ministry has stopped issuing declarations, the court must infer the position from conduct: whether an ambassador has been kept in place, whether the authority's diplomats have been accepted, and whether dealings continue. The political judgment moves from the executive to the judge, who is worse placed to make it.

6. Explain de facto and de jure recognition of governments, and give a consequence of the difference.

De facto recognition acknowledges that an authority is in effective control while reserving judgment on the lawfulness or permanence of its position; it is provisional and can be withdrawn without great difficulty. De jure recognition acknowledges the authority as the lawful government, is intended to be permanent, and is withdrawn only in grave circumstances. One consequence is that a State may recognise one authority de jure and another de facto at the same time in the same State, which happens during and after civil wars; another is that de jure recognition generally carries fuller consequences before the recognising State's courts, including immunity and the retroactive validation of the authority's acts.

Conditional and unconditional

Conditional recognition is granted on terms: that the new State protect minorities, observe religious freedom, accept certain boundaries, or assume a share of the predecessor's debts.

The orthodox view is that once recognition has been given it takes effect, and the breach of a condition is a breach of the undertaking rather than a ground for treating the recognition as void. So the condition creates an obligation and does not suspend the recognition.

Individual and collective

Individual recognition is what usually happens: each State decides for itself.

Collective recognition is recognition by a group of States acting together, or through an international organisation. It has been proposed repeatedly and has never been institutionalised. As explained in [Recognition of States], admission to the United Nations under article 4 is not collective recognition, and a Member does not thereby recognise every other Member.

Something close to it happens in practice when a group of States announces common criteria and then recognises together, which is how several new European States were recognised in the early 1990s.

Premature recognition

Recognition given before the entity satisfies the criteria, most obviously while the parent State is still fighting to suppress a secession.

Premature recognition is a wrongful act, an unlawful intervention in the affairs of the parent State, and the parent State may treat it as such. It is important because it shows recognition is not a purely discretionary act: the discretion runs out where the criteria are plainly not satisfied.

Withdrawal of recognition

Recognition of a government may be withdrawn when the authority loses control, and in practice it is more often allowed to lapse than formally revoked.

Recognition of a State is withdrawn only if the State ceases to exist, by absorption or merger or complete disintegration. It is not a means of expressing displeasure, and States that quarrel break off diplomatic relations instead, which is a different thing.

The legal effects

This is where the topic earns its place in a syllabus, because these are the consequences that decide real cases.

On the international plane

Diplomatic relations become possible, though they do not follow automatically: recognition and the exchange of ambassadors are separate steps.

Bilateral treaties become possible, since a State will not conclude one with an entity it does not accept as a State.

Claims may be made and received between the two, and the recognised entity may bring an international claim against the recognising State.

In the recognising State's own courts

Four consequences, and these are the ones examiners ask about.

Standing to sue. A recognised State or government may bring proceedings in the recognising State's courts. An unrecognised one generally cannot, because the court will not accept that it exists as a legal person for that purpose.

Sovereign immunity. A recognised State and its property are entitled to immunity, on the principles in [Immunity from Jurisdiction]. An unrecognised entity is not.

Validity of legislative and executive acts. The courts of the recognising State give effect to the laws and official acts of a recognised State or government within its territory: its marriages, its company registrations, its judgments, its transfers of property. Those of an unrecognised authority are, in strict theory, a nullity, which produces the injustice noted below.

Title to property abroad. Recognition, and particularly de jure recognition, entitles the government to the State's property in the recognising State, including bank accounts and embassy premises, and this is what the litigation is usually really about.

Retroactivity. De jure recognition is generally treated as dating back to the moment the authority actually came to power, so that acts done before recognition are validated. That doctrine exists precisely to reduce the injustice of treating a whole population's transactions as a nullity.

A worked example

The authorities of Ruritania, an entity that broke away from State W four years ago, sue in the courts of State X to recover a consignment of goods. Ruritania controls its territory, but State X has not recognised it. Meanwhile a couple married under Ruritanian law seek to have their marriage recognised in State X, and State W claims Ruritania's bank account in a State X bank.

Can Ruritania sue? If State X has not recognised it, generally no. The court will not treat it as a legal person entitled to appear.

Is Ruritania entitled to immunity if sued? No, on the same reasoning, which cuts the other way and is an argument the unrecognised entity will not make.

Is the marriage valid? In strict theory the acts of an unrecognised authority are a nullity in State X, so the marriage would not be recognised. That is a harsh result for people who had no choice in the matter, and it is the reason courts and legislatures have found ways round: recognising private-law acts of daily life even where the authority is not recognised, and applying the retroactivity doctrine as soon as recognition comes.

Who gets the bank account? State W, while it is the recognised claimant. If State X later recognises Ruritania de jure, the recognition relates back to the date Ruritania came to power, and Ruritania's title is treated as having existed throughout.

Change the facts: State X recognised Ruritania two years ago, while State W was still fighting to suppress the secession and Ruritania controlled only part of its claimed territory. That is premature recognition, and it is a wrongful act, an unlawful intervention in State W's affairs, of which State W may complain.

Distinctions table

What it does NOT mean

Recognition is not the same as diplomatic relations. Breaking off relations does not withdraw recognition, and recognition does not require an exchange of ambassadors.

Dealing with an entity is not recognising it. Negotiation, trade, membership of the same organisation and participation in the same multilateral treaty do not imply recognition, and States often say so expressly.

Conditional recognition is not conditional. Once given it takes effect, and breach of the condition is a breach of an undertaking, not a ground for treating the recognition as never given.

Withdrawal of recognition of a government is not withdrawal of recognition of the State.

Limits and criticism

The consequences fall on private people who had no say. Whether a marriage, a birth certificate, a company registration or a judgment is valid should not depend on a foreign ministry's political judgment, and the devices courts use to avoid that result are patches on a doctrine that produces the wrong answer.

The de facto and de jure distinction is imprecise. It is a distinction between degrees of confidence dressed as a distinction of kind, and States use the labels inconsistently.

Non-recognition is used as a sanction, which the doctrine cannot really bear. Recognition is supposed to acknowledge facts, and using it to punish makes the legal question turn on unrelated policy, which is the criticism the Estrada doctrine was built on.

Quick revision

Modes. Express or implied; de facto or de jure; conditional or unconditional; individual or collective.

Implied by: a bilateral treaty, full diplomatic relations, a consular exequatur. Not implied by: a multilateral treaty, common membership of an organisation, negotiation, trade contacts, or a vote for admission to the United Nations.

De facto is provisional and readily withdrawn; de jure is final and generally retroactive to the date power was assumed.

Conditional recognition takes effect once given; breach of the condition is a breach of an undertaking.

Collective recognition has never been institutionalised; article 4 admission is not recognition.

Premature recognition is a wrongful act, an unlawful intervention in the parent State's affairs.

Effects internationally: diplomatic relations become possible, bilateral treaties become possible, claims may be exchanged.

Effects in the recognising State's courts: standing to sue, sovereign immunity, validity of legislative and executive acts, and title to State property abroad, with retroactivity on de jure recognition.

Test yourself

1. List the modes of recognition and say which acts imply it.

Recognition may be express or implied, de facto or de jure, conditional or unconditional, and individual or collective. The acts generally treated as implying recognition are the conclusion of a bilateral treaty, the establishment of full diplomatic relations, and the grant of a consular exequatur. Acts not treated as implying it include being parties to the same multilateral convention, common membership of an international organisation, negotiation, unofficial and trade contacts, and voting for admission to the United Nations. Recognition is not lightly implied, and States that wish to avoid it commonly declare that their dealings are not to be taken as recognition.

2. Distinguish de facto from de jure recognition and give three consequences.

De facto recognition acknowledges effective control provisionally, reserving judgment on lawfulness and permanence; de jure recognition acknowledges lawful status and is intended to be final. First, de facto recognition may be withdrawn if the control it acknowledged ceases, while de jure recognition is withdrawn only in grave circumstances. Second, de jure recognition is generally retroactive to the date the authority came to power, so that its earlier acts are validated, while de facto recognition is not, or not to the same extent. Third, full diplomatic relations and the right to claim the State's property in the recognising State ordinarily follow de jure recognition.

3. What are the effects of recognition in the courts of the recognising State?

Four. The recognised State or government may sue there, while an unrecognised one generally may not. It and its property are entitled to sovereign immunity. Its legislative and executive acts within its own territory are given effect, so its marriages, registrations, judgments and transfers of property are treated as valid. And it is entitled to the State's property within the recognising State, including bank accounts and diplomatic premises. De jure recognition operates retroactively to the date power was assumed, validating acts done before recognition was given.

4. Is premature recognition lawful?

No. Recognition given before the entity satisfies the criteria of statehood, in particular while the parent State is still actively resisting a secession, is a wrongful act amounting to an unlawful intervention in the parent State's internal affairs, of which the parent State may complain. It shows that recognition is not a wholly discretionary political act: the discretion is exercised on facts, and it runs out where the facts are plainly absent.

5. What is the effect of breaching a condition attached to recognition?

On the orthodox view, none upon the recognition itself. Recognition once given takes effect, and the condition creates an obligation whose breach is a breach of an undertaking, giving rise to a complaint or to a claim, but not to the treatment of the recognition as void or as never having been given. The point matters because conditions of this kind were common in the recognition of new States in Europe, and the question of what happens when they are not kept has repeatedly arisen.

6. Why do courts treat the acts of an unrecognised authority as valid in some circumstances?

Because the strict rule produces injustice to private people who had no choice in the matter. If the acts of an unrecognised authority were a nullity everywhere, then marriages, births, deaths, company registrations, land transfers and judgments in a whole territory would have no effect abroad, and the population would be punished for a political decision taken elsewhere. Courts therefore distinguish the authority's public and political acts from the private-law acts of daily life, giving effect to the latter, and the doctrine that de jure recognition relates back to the date power was assumed removes the problem entirely once recognition is finally given.