Collective Security in International Relations

Collective security is the most ambitious organising idea in the history of international institutions and, by its own standard, the least successful: it has never once operated as its designers intended. Yet the League, the UN Charter, peacekeeping, the Responsibility to Protect and every proposal for Security Council reform are arguments conducted inside its vocabulary. The interesting question is not why it failed, but why states keep rebuilding an idea whose central requirement they have never been willing to meet.

The Idea and What It Asserts

• Collective security is the principle that an attack on any one member of the international community is an attack on all, and that all are obliged to respond collectively against the aggressor — a public guarantee replacing the private management of security.
  • The classic formulation is “one for all and all for one”: every state renounces the private use of force and receives the guarantee of the whole in exchange.
  • It works as an insurance system: each member pays a premium in the form of an advance commitment to act, and each is covered against aggression by any other.
  • It does not promise aggression will not occur, only that it will be met by preponderant force — a standing deterrent rather than a defence.
• The trigger is aggression as such, not aggression by a named enemy — what separates collective security from everything resembling it.
  • The obligation attaches to an act, not an identity; whoever commits the act becomes the target, and the system has no view in advance about who that will be.
  • There is therefore no target state, no war plan and no integrated command, because none can be prepared against an enemy unknown until the breach.

The Inside-Out Logic

Collective security faces inward. An alliance is built against a threat external to it; a collective security system treats the threat as arising from within its own membership. The organisation is at once a club and a police force over its members, and its enforcement is always directed at one of its own — which is the source of both its moral appeal and its practical impossibility.

• The threat comes from inside the system. Every member is a potential aggressor and a potential victim, which is why membership must be near universal.
• The peace is indivisible: aggression anywhere injures the legal order everywhere, so no member may treat a distant war as its own affair alone.
• Neutrality is renounced. There is no third position between aggressor and collectivity; a state that stands aside has already defected.

“There must be, not a balance of power, but a community of power; not organized rivalries, but an organized common peace.” — Woodrow Wilson

The Intellectual Lineage

• The peace-plan tradition of early modern Europe supplied the architecture long before the politics existed.
  • The Grand Design attributed to the Duke of Sully and Henry IV proposed a Christian republic of European states with a common council and a joint force to coerce a peace-breaker.
  • The Abbé de Saint-Pierre proposed a permanent congress of sovereigns guaranteeing existing frontiers, with collective coercion of any prince who violated them.
  • Jeremy Bentham proposed a common court of judicature, arguing that public opinion rather than armies would compel compliance — an early statement of the sanctions logic.
  • Kant’s Perpetual Peace proposed a federation of free states, republics bound by a pacific union renouncing war among themselves — the philosophical ancestor of the Covenant, though Kant refused a world state.
• The Concert of Europe was the proto-form. After 1815 the great powers consulted at congresses, treated the European settlement as a common possession and intervened jointly to preserve it.
  • It shows the instinct — great powers acting in concert to manage the system rather than divide it — while remaining an oligarchy of the strong, not a universal guarantee.
  • Its collapse in 1914 made the demand for a formalised, universal system irresistible.
• Woodrow Wilson is the modern author of the concept, and he offered it explicitly as the replacement for the balance of power.
  • Wilson blamed the alliance system and the balance of power for turning a Balkan quarrel into a world war: interlocking commitments made a local dispute automatically general.
  • The fourteenth of the Fourteen Points called for a general association of nations guaranteeing political independence and territorial integrity to great and small states alike.
  • His charge was that the balance of power gives protection against measles while creating the conditions for plague.
  • Nehru made the same objection in calling the balance of power a nervous state of peace, in which the absence of war is only the interval between wars.

The Assumptions the System Requires

Collective security is not a loose aspiration but a demanding set of conditions, all of which must hold at once. No international system has ever satisfied more than three or four of them simultaneously, and the two hardest have never been met at all.

Requirement	What it demands	Why it is hard
Near-universal membership	Every state of consequence inside the system, so no aggressor sits outside the guarantee	The League never had the United States, and the Soviet Union only from 1934 to 1939
Agreed definition of aggression	A shared, applicable test of who struck first and who is the victim	Identifying an aggressor is a political act, not a factual finding
Overwhelming preponderance	The collectivity must so far exceed any aggressor that resistance is futile	Untrue against a great power, and untrue whenever the collectivity is divided
Willingness to act against a friend	Every member must be prepared to coerce an ally, a patron, a client — or itself	Never once honoured against a great power
Indivisibility of peace	Distance and disinterest are no defence; every breach concerns everyone	States discriminate ruthlessly between wars that touch them and those that do not
Automaticity	The response follows the breach without fresh political bargaining	Both the Covenant and the Charter left the determination to a political organ
Renunciation of neutrality	No member may stand aside once aggression is determined	Non-alignment, abstention and quiet trade with the aggressor are the norm
Particular yields to general interest	The obligation to the system outranks the national interest of the moment	The realist objection in its purest form
Rough equality of commitment	Burdens fall on all, not on whichever member happens to care	Produces a chronic free-rider problem

Requirement by Requirement, Why Each Has Failed

• Universality never existed, and the exceptions were always the states that mattered: the United States never ratified the Covenant it had drafted, the Soviet Union was outside it for most of its life, and the revisionist powers withdrew once it began to constrain them.
  • UN membership is genuinely near-universal at 193 states — and the system still does not work, which shows universality was never the binding constraint.
• Agreement on what constitutes aggression has proved unattainable in the cases that matter.
  • The Assembly adopted a definition of aggression in 1974 and it entered criminal law through the Kampala amendments of 2010, but neither binds the Council’s political determination.
  • Every hard case is contested: pre-emption presented as self-defence, intervention as protection of nationals, proxy war as civil war.
• Preponderance holds against the weak and dissolves against the strong, making collective security a mechanism that operates only downward.
• The requirement to act against a friend has never been met.
  • Britain and France would not coerce Italy over Abyssinia because Mussolini was wanted as a counterweight to Hitler.
  • Every permanent member has since vetoed to shield itself or a client, and the pattern is invariant across ideology.
• Automaticity was deliberately abandoned in drafting. The Covenant qualified its obligations by procedure and unanimity; the Charter routed everything through a discretionary determination under Article 39. Both chose political control over automatic enforcement, removing what made the idea a guarantee.
• The general obligation has consistently lost to the particular interest. The design assumes international peace is the aim of states, when the aim of states is the national interest.

“Should any Member of the League resort to war in disregard of its covenants … it shall ipso facto be deemed to have committed an act of war against all other Members of the League.” — Covenant of the League of Nations, Article 16

The Covenant’s language could hardly be stronger, and its strength is the problem: an obligation that severe was never going to be performed, and its non-performance discredited not the League alone but the idea.

Collective Security Against the Balance of Power

Collective security was designed as the replacement for the balance of power and is best understood against it. The two are rival answers to one question — how do you stop power being abused? — and they agree that only power restrains power before diverging entirely about how that power is assembled. For balancing mechanics, alliances, polarity and BoP techniques see the companion article Balance of Power.

What the Two Share

• Both are methods of managing power, resting on the view that greater power is the antidote to power; neither imagines peace secured by law or goodwill.
• Both treat state sovereignty as sacred: a collective security organisation defends its members’ sovereignty, it does not supersede it.
• Both treat war as a legitimate instrument for defending that sovereignty. Collective security is not pacifism — it licenses war against an aggressor and makes joining it an obligation.

Where They Diverge

Dimension	Balance of power	Collective security
Setting	Operates in anarchy, requiring no institution	Requires an organisation — the League, the UN
Character	Ad hoc, improvised case by case	Institutionalised, laid down in advance in a treaty
Predictability	Uncertain — a balancing coalition may or may not form	Theoretically certain — the response is stipulated
The enemy	A specific rising or threatening power	Any aggressor, not identified in advance
Direction of the threat	External to the coalition	Internal to the system
Primary beneficiary	Great powers, who do the balancing	Small and poor states, who are guaranteed without arming
Effect on armament	Drives arms races and alliance-building	Designed to reduce both

• The uncertainty of balancing is its central practical defect, and Napoleon is the illustration: no counterbalancing coalition formed automatically, and it took a decade and seven coalitions to restore the balance. Balancing is a tendency, not a mechanism, and it may arrive too late.
• Collective security is highly attractive to small and poor states, offering the protection of the whole without the cost of armament or the political price of alliance — which is why they are its most consistent defenders and great powers its most consistent saboteurs.
• Collective security can be read as an institutionalised balance of power — the same aggregation of superior force against a disturber, with membership fixed in advance and the coalition assembled through a standing organisation.

Complementary and Antagonistic at the Same Time

• They are complementary: the collective security machinery is built on great-power preponderance and cannot function without it.
  • Chapter VII enforcement presupposes that the permanent members together command overwhelming force. The UN has no army; the collectivity’s preponderance is the great powers’ preponderance, borrowed.
  • The veto is a balance-of-power device inserted into a collective security charter — it records that the five could not be coerced, and converts the Council into a great-power concert with a universal membership attached.
• They are antagonistic: each denies the other’s founding premise.
  • The balance of power treats security as a private good, produced by each state through armament and alliance; collective security treats it as a public good, produced once for everyone.
  • Balancing requires flexible, reversible commitments and the freedom to change sides; collective security requires fixed, universal and irreversible commitments and forbids sides altogether.
  • Balancing treats neutrality as a legitimate option; collective security treats it as defection.
  • Success on one measure is failure on the other: where balancing works, no state trusts the guarantee; where the guarantee is trusted, no state balances.

Collective security did not replace the balance of power. It was built on top of it, and the veto is where the older logic surfaces through the newer one.

Collective Security Against Collective Defence

The commonest error in this field is to treat these as one thing, usually by calling NATO a collective security organisation. They share a slogan — an attack on one is an attack on all — and almost nothing else. NATO is a military alliance, not a collective security organisation, and the difference is the whole structure of each concept.

Dimension	Collective security	Collective defence
Membership	Universal or near-universal	Selective — a chosen alliance
Source of threat	From inside the system: any member that commits aggression	From outside the alliance
The enemy	Undefined in advance; identified only when aggression occurs	Known in advance, which is why the alliance exists
Obligation	Against any aggressor, including a friend or oneself	Against an external attacker only
Advance planning	Impossible — no target, no theatre	Essential — war plans, force goals, exercises
Integrated command	Absent; forces assembled after the breach	Present and standing
Legal basis	Chapter VII of the UN Charter; Covenant Articles 10 and 16	Article 51 of the Charter plus the alliance treaty
Effect on armament	Intended to reduce arms racing	Tends to stimulate it and to provoke counter-alliance
Examples	League of Nations, United Nations	NATO Article 5, the former Warsaw Pact, ANZUS

• Advance planning is the sharpest practical test. An alliance knows whom it will fight and can build an integrated command, pre-position stocks and rehearse. A collective security system cannot: to prepare against a named enemy is to abandon the premise that the enemy is whichever member offends.
• Article 5 of the North Atlantic Treaty has been invoked exactly once, after 11 September 2001 — and the attacker was a non-state network, not the state adversary the alliance was built against.
  • NATO has since grown to 32 members and raised its pledge to 5 per cent of GDP by 2035, but its obligation still runs only to its members and only against external attack.

“Nothing in the present Charter shall impair the inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations, until the Security Council has taken measures necessary to maintain international peace and security.” — UN Charter, Article 51

The Indian and the American Position

This is a genuine disagreement about whether collective defence supports or corrodes collective security, and both positions are coherent.

• India opposed collective defence pacts on the ground that they undermine faith in collective security.
  • The objection was pressed hardest against SEATO and CENTO, which brought extra-regional military commitments into South and West Asia and drew Pakistan into a great-power alignment.
  • The argument: if states may contract out of the universal guarantee into a private one they will, leaving the universal system only the cases nobody has contracted for. The alliance is a vote of no confidence in the Charter, cast by the states best able to make it work.
  • It also imports the bloc logic collective security was invented to abolish, reproducing the alignment structure Wilson blamed for 1914.
  • Behind the legal argument lay a political one: pacts subordinate a smaller member’s foreign policy to its patron, which is why non-alignment and support for collective security were the same policy stated twice.
• The United States sees no contradiction, treating collective defence as the practical way to operationalise collective security.
  • Article 51 preserves the inherent right of individual and collective self-defence, so an alliance is not a departure from the Charter but an exercise of a right the Charter expressly saves.
  • Where the Council is paralysed, an alliance is the only body capable of the preponderant response; the alternative is not the Charter working but nothing happening.
  • Alliances supply the integrated command, interoperability and readiness the UN has never possessed, and every large enforcement operation has been executed by an alliance or a coalition.
• The disagreement is unresolved, because both are right: alliances do supply the capability the UN lacks, and they do erode the universality on which the UN’s claim rests.

Two Further Distinctions: Cooperative and Common Security

Concept	Core idea	Referent	Method
Collective security	Aggression by any member is met by all	The state and the legal order	Enforcement after a breach
Collective defence	Attack on a member is met by the alliance	The alliance’s members	Deterrence and defence against an external enemy
Common security	Security with the adversary, not against it	The relationship between adversaries	Mutual restraint, reassurance, arms control
Cooperative security	Habits of consultation, transparency and confidence-building among all	The system’s stability	Prevention through dialogue and inclusion, not enforcement

• Common security is the doctrine of the Palme Commission, which held that in the nuclear age security cannot be had against an adversary but only jointly with it, since a gain at the other’s expense destabilises. The OSCE is its fullest expression.
• Cooperative security is inclusive, consultative and preventive, building confidence so aggression does not arise — the ASEAN Regional Forum is the model. Dispensing with enforcement makes it the easiest form to accept and the weakest against a determined aggressor.

The League of Nations Record

The Covenant Machinery

• Article 10 was the heart of the guarantee. Members undertook to respect and preserve against external aggression the territorial integrity and existing political independence of all members, with the Council to advise on the means.
  • It made the territorial status quo a collective possession, and the League a defender of the 1919 settlement — which is why revisionist powers read it as an instrument of the victors. The word advise is where the guarantee leaked: the Council recommended, members decided.
• Article 11 declared any war or threat of war a matter of concern to the whole League, whether or not a member was affected — the textual statement of the indivisibility of peace.
• Article 12 imposed a cooling-off obligation: disputes to arbitration, judicial settlement or Council inquiry, and no resort to war until three months after the award or report.
  • The Covenant therefore did not outlaw war; it delayed and regulated it, leaving a gap a patient aggressor could walk through. The Kellogg-Briand Pact of 1928 tried to close that gap by renouncing war as an instrument of national policy.
• Article 15 governed disputes likely to lead to rupture: the Council reported, and if the report was unanimous excluding the parties, members agreed not to go to war with a complying party.
• Article 16 was the sanctions article. A member resorting to war in breach of the Covenant was deemed ipso facto to have committed an act of war against all other members, who were to sever trade and financial relations; the Council was to recommend what forces members should contribute.
  • Sanctions were automatic in text and discretionary in fact, each member judging its own obligation, and military measures were only ever recommendations — no standing force, no obligation to supply one, and the Covenant’s largest structural defect.

What the League Actually Achieved

Its successes were real and share one diagnostic feature: in every case, no great power’s interest lay with the disturber.

• The Åland Islands, 1921 — a Finnish-Swedish dispute over an archipelago with a Swedish-speaking population, settled by awarding sovereignty to Finland with guaranteed autonomy, demilitarisation and language rights; it has held for a century.
• Upper Silesia, 1921 — after a contested plebiscite and communal violence, the League partitioned it between Germany and Poland under a fifteen-year minority-protection regime.
• The Greek-Bulgarian border incident of 1925 — Greek forces crossed into Bulgaria; the Council ordered withdrawal within sixty hours, verified it and made Greece pay compensation. Both parties were small; both complied.
• Also the Mosul question, Memel and the Saar plebiscite.

The Failures That Killed It

• Manchuria, 1931 — Japan manufactured the Mukden incident, seized Manchuria and installed the puppet state of Manchukuo.
  • The Lytton Report of 1932 found Japan’s action was not self-defence and Manchukuo not a genuine independence movement, while acknowledging Japanese grievances and proposing autonomy under Chinese rule.
  • The Assembly adopted it in February 1933; Japan walked out and left the League, and no sanctions followed — a great power could defy the League, be found guilty in a formal report, and suffer nothing.
• Abyssinia, 1935 — the case that destroyed the League’s credibility. Italy invaded Ethiopia, a member state whose emperor, Haile Selassie, addressed the Assembly in person to demand the collective guarantee.
  • The League did find Italy the aggressor and did impose Article 16 sanctions — the only time the article was invoked against a great power — but engineered them to fail: oil was excluded, the Suez Canal stayed open to Italian troopships, and the non-member United States kept trading.
  • The Hoare-Laval pact of December 1935, a secret Franco-British plan to hand Italy most of Ethiopia, leaked and forced both foreign ministers out. The two states charged with enforcing the Covenant had been caught negotiating the partition of the victim.
  • Sanctions were lifted in July 1936 once the conquest was complete, and Haile Selassie’s warning that the League’s own existence was in issue was vindicated within four years.
• The Chaco War, 1932-35 — Bolivia and Paraguay, both members, fought Latin America’s bloodiest war of the century; the arms embargo came late and the settlement was brokered outside the League.
• The Rhineland, 1936 — remilitarisation in breach of Versailles and Locarno drew condemnation and nothing else; Anschluss, Munich, Prague and Poland followed with the League a bystander.

Why It Failed: The Structural Reasons

• The United States never joined the organisation its President had designed. The Senate refused ratification, chiefly over Article 10’s open-ended commitment, and the absence of the largest economy stripped sanctions of their force.
• The Soviet Union was outside for most of the period — excluded until 1934, expelled in 1939 after the attack on Finland — so two of the four greatest powers stood outside the League for most of its life.
• Britain and France had no real faith in it, treating it as an adjunct to their own diplomacy; the Hoare-Laval pact is the proof.
• Appeasement of the fascist powers followed from a judgement that communism was the larger threat, which made Italy and even Germany potential partners rather than aggressors to be coerced.
• The unanimity rule in the Council gave every member a veto. Substantive decisions required unanimity of members present, parties excepted. The Covenant’s defect was not too few vetoes but too many.
• There was no standing force, no obligation to supply one and no planning machinery, so even a Council that agreed had nothing to act with.
• Sanctions were weak in design and weaker in use — self-assessed, incomplete, evadable through non-members, and undercut by the wish to keep the sanctioned state’s goodwill.
• The deepest objection is conceptual: the system assumed international peace is the aim of states, when the aim of states is the national interest.

The League did not fail because its rules were badly drafted. It failed because its rules required states to do something they had never agreed to do.

The United Nations Charter Machinery

The Charter’s drafters were reformers of the League, not inventors of a new idea. Their remedy for the Covenant’s weakness was to concentrate enforcement in a small executive organ with binding powers and a great-power core. Article 1(1) makes peace and security the organisation’s first purpose; Article 2(4) prohibits the threat or use of force against any state’s territorial integrity or political independence.

Chapter VII: Action With Respect to Threats to the Peace, Breaches of the Peace and Acts of Aggression

Provision	What it does	What became of it
Article 39	The Council determines a threat to the peace, breach of the peace or act of aggression and decides what measures to take	The gateway to all enforcement; wholly discretionary and political
Article 41	Measures not involving armed force — economic, communications and diplomatic severance	The most-used power: sanctions, embargoes, tribunals, listing regimes
Article 42	Action by air, sea or land forces where Article 41 is or proves inadequate	Never used as designed; delegated to willing states
Articles 43-47	Special agreements placing forces at the Council’s disposal, with a Military Staff Committee of the P5 chiefs of staff to direct them	Never operationalised; no special agreement has ever been concluded
Article 51	Preserves the inherent right of individual or collective self-defence until the Council acts	The basis of every alliance, and the widest loophole
Article 27(3)	Nine affirmative votes including the concurring votes of the permanent members	The veto
Chapter VIII	Regional arrangements for local disputes; no enforcement without Council authorisation	How the AU, ECOWAS, NATO and coalitions now do most enforcement

• Articles 43 to 47 are the missing spine of the system. The Charter assumed forces on call under professional military direction; the Cold War arrived before any special agreement was negotiated. Every enforcement operation since has been authorisation by delegation — the Council licenses willing states and then has no command over what they do.
• Article 51 creates the loophole that swallows the prohibition. Self-defence is inherent and exercisable without permission, and its limits are all contested — what counts as an armed attack, whether anticipatory action qualifies, whether it reaches non-state attackers on an unwilling host’s territory.
  • A state that wishes to use force will frame it as self-defence, and the Council rarely gets to rule.
• The veto was a condition of the organisation’s existence, not an afterthought: the great powers would not join a system able to coerce them, and the alternative — the League’s unanimity of all Council members — was worse.
  • Restricting the veto to five is a real improvement; it also writes great-power immunity into a system whose premise is that no one is immune.
• Chapter VIII has become the operational reality, with regional organisations conducting most enforcement and authorisation sought before, during or after the fact.

The Record: The Two Cases That Are Cited

• Korea, 1950 — the clearest case, with an honest qualification.
  • North Korea invaded the South on 25 June 1950; the Council determined a breach of the peace and recommended that members furnish assistance and place forces under a unified command under the United States, using the UN flag.
  • It happened only because the Soviet representative was absent, the USSR having boycotted the Council over its refusal to seat the People’s Republic of China. Once Moscow grasped what its absence had permitted it returned in August 1950 and vetoed, and no further authorising resolution was possible.
  • In substance it was an American-led war under a UN flag, with sixteen states contributing combat forces but the command, strategy and bulk of the forces American, and the field commander reporting to Washington rather than the Council.
• The Gulf War, 1990-91 — the genuinely closest instance.
  • Iraq invaded Kuwait on 2 August 1990. The Council condemned it the same day, imposed Article 41 sanctions within four days, declared the annexation null and void, and in Resolution 678 of 29 November 1990 authorised states cooperating with Kuwait to use all necessary means unless Iraq withdrew by 15 January 1991.
  • More than thirty states contributed forces, including Arab states; the aggression was reversed and the victim’s sovereignty restored.
  • The conditions were exceptional: the Cold War had just ended, the Soviet Union acquiesced, China abstained, and the aggression was an outright annexation.
  • The criticism is that this was authorisation by delegation, not collective security under UN command: the Council licensed a coalition and controlled neither the war’s conduct nor its termination.
  • The same “all necessary means” formula later stretched to a decade of containment and, in 2003, to an argument for invasion the Council largely rejected.

The Cases Where the Council Was Paralysed

The pattern is invariable: wherever a permanent member is a party or the patron of a party, the machinery does not move.

• Hungary, 1956 — the Soviet veto blocked the Council after Soviet tanks crushed the uprising, and the matter went to the Assembly; Suez, 1956 — Britain and France vetoed resolutions aimed at their own invasion of Egypt, the only crisis in which two permanent members vetoed to protect themselves.
• Vietnam — never seriously before the Council, the party being a permanent member; Afghanistan, 1979 — the Soviet invasion was vetoed out and condemned only by the Assembly.
• Kosovo, 1999 — NATO bombed Yugoslavia for eleven weeks without Security Council authorisation, anticipating a Russian and Chinese veto; widely defended as necessary and indefensible as lawful.
• Iraq, 2003 — invaded without a fresh resolution, on a contested reading of earlier ones, against three permanent members’ declared opposition. The Council’s authority over the use of force stopped being assumed here.
• Syria, from 2011 — repeated vetoes shielded the government from referral, sanction and accountability for chemical weapons.
• Ukraine, since 2022 — a permanent member is the aggressor and votes on its own aggression, so the Council cannot even make an Article 39 determination.
• Gaza, since 2023 — repeated vetoes on ceasefire texts, almost all cast by one permanent member protecting a close ally.

Uniting for Peace

• The Acheson Plan, adopted as General Assembly Resolution 377(V) on 3 November 1950, was the American answer to the Soviet return to the Council after Korea — a route around the veto.
• What it provides.
  • Where the Council, because of a lack of unanimity among the permanent members, fails to exercise its primary responsibility over a threat to the peace, breach of the peace or act of aggression, the Assembly shall consider the matter immediately.
  • The Assembly may make recommendations for collective measures, including the use of armed force, to restore international peace and security.
  • It may meet in emergency special session within 24 hours, convened by any nine members of the Council — a procedural vote immune to the veto — or by a majority of the membership, and recommends by a two-thirds majority.
  • It also created a Peace Observation Commission and a Collective Measures Committee and invited members to earmark national units; that machinery was never built.
• Its use. There have been eleven emergency special sessions, and the roll call is the Cold War and its aftermath.
  • Suez, 1956 — the first and most consequential: with Britain and France vetoing, the Assembly demanded a ceasefire and authorised the first armed peacekeeping force.
  • Then Hungary 1956; Lebanon and Jordan 1958; the Congo 1960; the Middle East 1967; Afghanistan 1980; Palestine from 1980; Namibia 1981; the occupied Arab territories 1982.
  • The tenth session, on illegal Israeli actions in occupied East Jerusalem and the occupied Palestinian territory, has been resumed repeatedly since 1997 and has carried Assembly action over Gaza since 2023.
  • The eleventh session, on Ukraine, was convened in 2022 on a Council procedural vote Russia could not veto, and produced large majorities demanding withdrawal, condemning annexation and addressing reparation — the mechanism’s most prominent revival in decades.
• The objections are serious and unanswered.
  • The Soviet and Russian position is that the resolution is ultra vires: it is not in the Charter, and Article 24 confers primary responsibility on the Council, so a procedural device cannot move a treaty power between organs.
  • Article 11(2) requires the Assembly to refer to the Council any question on which action is necessary — a textual argument that enforcement is not the Assembly’s to authorise.
  • The Assembly can only recommend, not bind. A recommendation to use force creates no obligation and commands no forces, so what it produces is permission rather than compulsion.
  • The International Court of Justice, in Certain Expenses, upheld the Assembly’s power to recommend and to organise non-coercive operations, securing peacekeeping’s legality without settling the enforcement question.
  • It has also been used selectively by whichever bloc held the Assembly majority, which is why the same states have defended and attacked it at different times.
• The honest assessment is that it is an instrument of legitimacy rather than enforcement. No enforcement action has ever been mounted on its authority alone.
  • What it does is deny an aggressor the shelter of a vetoed Council: a recorded global majority, an authoritative finding of breach, and a basis for sanctions adopted nationally.

Uniting for Peace cannot compel anyone. What it can do is make a vetoed silence audible.

Peacekeeping as the Substitute

• Peacekeeping was invented because collective security did not work. With the Council deadlocked, Dag Hammarskjöld devised a UN military deployment that required no enforcement at all: lightly armed, consent-based, interposed between parties who had already stopped fighting.
  • Lester Pearson, Canada’s foreign minister, proposed the first armed force during the Suez crisis and received the Nobel Peace Prize for it; UNEF I, authorised by the Assembly under Uniting for Peace in 1956, was the first of its kind.
  • The antecedents are the unarmed observer missions — UNTSO in the Middle East from 1948 and UNMOGIP on the India-Pakistan ceasefire line from 1949 — both still deployed.
• Peacekeeping appears nowhere in the Charter. Hammarskjöld called it “Chapter Six and a Half” — more than pacific settlement under Chapter VI, less than enforcement under Chapter VII. It is an extra-constitutional growth, legitimated by practice rather than by text.
• The three principles are what distinguish it from enforcement.
  • Consent of the parties — peacekeepers deploy by invitation and can be asked to leave, as UNEF I was in 1967, just before the Six-Day War.
  • Impartiality — the force takes no side, meaning it holds all parties to the mandate rather than treating them as morally equivalent.
  • Non-use of force except in self-defence and defence of the mandate — the last clause has expanded enormously and is where the doctrine now strains.

Three Ways of Using Force Under the UN Flag

	Collective security	Peacekeeping	Peace enforcement
Legal basis	Chapter VII, Articles 39-42	No Charter basis — “Chapter Six and a Half”	Chapter VII authorisation
Consent of the parties	Not required	Required and foundational	Not required
Target	The aggressor, named by the Council	No target; the force is interposed	A named spoiler or party
Impartiality	None — the system takes the victim’s side	Central	Compromised by design
Use of force	Preponderant, to defeat aggression	Self-defence and defence of the mandate only	Offensive operations authorised
Trigger	Aggression across a border	A ceasefire or agreement already in place	Continuing violence or atrocity

The Generations and the Doctrine They Produced

• First generation: traditional interpositional peacekeeping — monitoring ceasefires and buffer zones between consenting states, success measured by the absence of incidents: UNEF, UNDOF on the Golan, UNFICYP in Cyprus.
• Second generation: multidimensional peacekeeping after the Cold War, extending mandates to disarmament and demobilisation, elections, policing, human rights monitoring, refugee return and civil administration — Namibia, Cambodia, Mozambique, El Salvador.
• Third generation: robust mandates and peace enforcement, deploying where there was no peace to keep with Chapter VII authority to use force beyond self-defence.
  • Protection of civilians became a standard mandate task from UNAMSIL in Sierra Leone in 1999 and now covers most deployed personnel.
  • MONUSCO’s Force Intervention Brigade, created in 2013, was the first UN force mandated to conduct offensive operations against armed groups; it broke the M23 rebellion that year and has struggled since. It sits uneasily with impartiality and consent, and made UN troops a party to a war.
• The failures that shaped doctrine were all one failure: mandates and means that did not match the situation.
  • Somalia, 1993 — creep from famine relief to disarming a faction, the deaths of Pakistani and American soldiers, and a Western retreat from ground deployment.
  • Rwanda, 1994 — a small force, an explicit warning ignored, a refusal to reinforce or to use the word genocide, and the force cut while roughly 800,000 people were killed.
  • Srebrenica, 1995 — a UN-declared safe area held by a lightly armed Dutch battalion, air support withheld, and around 8,000 Bosniak men and boys murdered when it fell.
• The doctrinal response came in four documents.
  • The Brahimi Report, 2000 — mandates must be clear, credible and achievable; forces must be able to defend the mandate and to distinguish victim from aggressor.
  • The Capstone Doctrine, 2008 — the first consolidated statement of principles, restating consent, impartiality and minimum force while conceding that robust mandates operate at their edge.
  • The HIPPO report, 2015 — the High-Level Independent Panel urged the primacy of politics: peacekeeping serves a political settlement and cannot substitute for one.
  • Action for Peacekeeping (2018) and A4P+ (2021) — mutual commitments on performance, safety, women’s participation and political solutions.

The Current Crisis

• Peacekeeping is in its deepest contraction in a quarter of a century. At 31 December 2025 there were 58 multilateral peace operations across 34 countries, 18 of them UN-led, deploying 78,633 international personnel.
  • That is a 49 per cent fall from 2016, the lowest since at least 2000, with a 17 per cent drop in 2025 alone.
• The cause is financial. By July 2025 peacekeeping faced a shortfall of about $2 billion against a $5.6 billion budget, driven by arrears among the largest assessed contributors.
  • The response was a 15 per cent cut in expenditure, repatriation of a quarter of all uniformed personnel from autumn 2025, and a 2026/27 budget reduced by nearly half a billion dollars.
  • Missions shrink not because conflicts ended but because the money stopped — the sharpest illustration of how conditional the collective commitment is.
• Four operations closed in 2025 — the security support mission in Haiti, the OSCE presence on Nagorno-Karabakh, the SADC mission in the Congo and the UN assistance mission in Iraq — and two were launched, AUSSOM in Somalia and the Gang Suppression Force in Haiti, both regional or ad hoc rather than UN-commanded.
• All ten largest troop contributors are from the Global South — Uganda, Nepal, Bangladesh, India, Rwanda, Ethiopia, Burundi, Kenya, Pakistan and Indonesia. The states that authorise operations and the states that staff them are disjoint: the Council writes mandates whose risks are borne by others.
• The Secretary-General’s 2026 review records the conditions around this — the largest number of armed conflicts since 1945, more battle deaths in five years than in the preceding twenty, and interstate conflicts doubling between 2024 and 2025. Demand rises exactly as supply collapses.

India’s Contribution

• India is the largest cumulative contributor of peacekeepers in UN history, with more than 290,000 personnel across more than 50 missions since the 1950s.
• Around 5,000 Indian personnel — troops, police and staff officers — are currently deployed across nine missions, and India ranked fourth among military contributors to multilateral peace operations in 2025 with about 4,090 troops, behind Uganda, Nepal and Bangladesh.
• Nearly 180 Indian peacekeepers have died in service — the highest toll of any contributor.
• The contribution began early and has been continuous — medical units to Korea from 1950, the custodian force after the Korean armistice, command in the Congo in the 1960s, then Gaza, Cambodia, Somalia, Rwanda, Sierra Leone, Lebanon, South Sudan and the Congo.
• India deployed the first all-female Formed Police Unit in UN history, to Liberia in 2007; over 150 Indian women peacekeepers currently serve across six missions, and the Liberian deployment is credited with a marked rise in local women’s enrolment in the police.
• The argument this supports is a strong one. If the Council’s machinery has been paralysed for most of the organisation’s life, then peacekeeping is the form in which the UN has actually done its core work — and India has done more of it than any other state.
  • The claim rests on this record with population, economy, military weight and the deficit of a Council with no permanent Asian member but China.
  • The paradox states itself: the state that has done most to keep the organisation relevant to its founding purpose has no permanent voice in the organ deciding where it acts.
• The counter-arguments belong here too: troop contribution is not strategic responsibility, contributors are reimbursed, and permanent membership turns on what the existing five will concede.

Humanitarian Intervention: R2P’s Predecessor Problem

• Humanitarian intervention is the use of force by one or more states against another to halt grave suffering of that state’s own population — the dominant frame of the 1990s, from northern Iraq and Somalia to Bosnia, Rwanda’s absent intervention and Kosovo.
• The legality-versus-legitimacy problem. Article 2(4) prohibits force, Article 51 permits it only in self-defence, Chapter VII only with authorisation. There is no fourth gateway for humanitarian purposes, so an unauthorised intervention is unlawful whatever its motive.
  • Kosovo, 1999 produced the defining formula: the Independent International Commission on Kosovo found NATO’s campaign “illegal but legitimate”.
  • The alternative reading is worse: if force is lawful whenever a state believes the cause good, Article 2(4) has no content, and only states with the capacity to intervene will ever exercise the privilege.
• The term itself became unusable: post-colonial states heard the civilising mission in it, and the selectivity of practice confirmed the suspicion. R2P was an attempt to escape it, shifting the vocabulary from a right of the intervener to a responsibility of the state and insisting on Council authorisation rather than working around it.

The Responsibility to Protect

Origins

• Francis Deng’s argument that sovereignty is responsibility is the conceptual origin. Working on internal displacement, Deng argued that sovereignty is a function rather than a licence: authority over a population is conditional on discharging the duty to protect it. It converts an absolute claim into a contingent one.
• Kofi Annan put the dilemma to the General Assembly directly in the Millennium Report, after Rwanda, Srebrenica and Kosovo.

“If humanitarian intervention is, indeed, an unacceptable assault on sovereignty, how should we respond to a Rwanda, to a Srebrenica — to gross and systematic violations of human rights that offend every precept of our common humanity?” — Kofi Annan

• The International Commission on Intervention and State Sovereignty, convened by Canada, answered in its report The Responsibility to Protect in 2001.
  • It reframed a right to intervene into a responsibility to protect, held first by the state and only residually by the international community.
  • It set out three responsibilities — to prevent, to react and to rebuild — prevention the most important and least practised.
  • It laid down precautionary criteria for force: just cause, right intention, last resort, proportional means, reasonable prospects and right authority, the last identified as the Security Council.
• The doctrine was adopted in narrowed form at the 2005 World Summit in paragraphs 138 and 139 of the Outcome Document, endorsed by the Council in Resolution 1674, and on the Assembly’s formal agenda since 2021.

The Three Pillars

• Pillar one: the state’s own responsibility to protect its population from the four crimes — uncontroversial, and a restatement of existing law.
• Pillar two: international assistance and capacity-building to help states meet it — the pillar that does most of the work and attracts almost no attention.
• Pillar three: timely and decisive collective response through the Council, including Chapter VII measures, where a state manifestly fails to protect its population. This is the contested pillar and the whole of the controversy.

“We are prepared to take collective action … through the Security Council … should peaceful means be inadequate and national authorities are manifestly failing to protect their populations.” — 2005 World Summit Outcome, paragraph 139

R2P and Collective Security: Similar but Different

They are cousins, not twins. Both accept that some conduct justifies a coercive collective response and both make the Security Council the gatekeeper. Everything else — what triggers them, whom they protect and what they may do — differs.

	Collective security	Responsibility to Protect
Referent object	The state and the inter-state legal order	The population, including against its own state
Trigger	Aggression — open-ended, any breach of the peace	Four crimes only: genocide, war crimes, ethnic cleansing, crimes against humanity
Direction of the threat	Across borders, state against state	Within borders, usually state against its own people
Goal	Restore the status quo ante and deter future aggression	Protect people; regime and territory are incidental
Relation to sovereignty	Sovereignty is sacred and defended	Sovereignty is conditional on responsibility
Primary method	Preponderant force against the aggressor	A continuum: prevention, assistance, capacity-building, force last
Obligation	Of members to each other under the Charter	A political commitment, not a treaty obligation; no binding duty to act
Authority	Security Council, Chapter VII	Security Council, Chapter VII — the point they share
Legal status	Charter law, binding	Emerging norm, contested in application

• The referent-object shift is the deepest difference. Collective security protects states from each other and is indifferent to what a government does inside its frontiers; R2P protects populations from their own governments, reaching precisely the conduct collective security was designed to leave alone.
• The scope difference is the sharpest. Collective security’s trigger, aggression, is open-ended; R2P’s is a closed list of four crimes, narrowed at the 2005 Summit from the broader ICISS formulation to reassure states that feared an open-ended licence to intervene.
• The method difference is the most misunderstood. R2P is overwhelmingly preventive and assistive, with force under pillar three the exception; collective security has no preventive limb at all.
• The convergence is real, and it is where both are weakest: both require the Council, and the Council is where a permanent member’s interest stops everything.

The Record

• Libya, 2011 — the first and only authorisation of force against a functioning state for protection purposes.
  • Resolution 1970 of 26 February 2011 invoked R2P expressly, imposing an arms embargo, targeted sanctions and a travel ban, and referring the situation to the International Criminal Court.
  • Resolution 1973 of 17 March 2011 authorised all necessary measures to protect civilians and a no-fly zone, passed 10-0 with five abstentions — Russia, China, India, Brazil and Germany.
  • The operation ended in regime change and the killing of Gaddafi, an outcome no one could locate in a civilian-protection mandate, and Libya collapsed into a decade of civil war and militia rule.
  • The consequence has been decisive. The abstaining states concluded that protection had been cover for a political objective, and the trust the next authorisation would need was destroyed.
• Côte d’Ivoire, 2011 — the less noticed success: Resolution 1975 authorised the UN mission to use all necessary means to protect civilians after Laurent Gbagbo refused to accept defeat. Kenya, 2008 is the preventive success, achieved by mediation with no force at all.
• The failures define its reputation.
  • Syria — a state manifestly failing and the Council veto-blocked through more than a decade of atrocity, including chemical-weapons use.
  • Myanmar — the expulsion of the Rohingya in 2017, met by a Council unable to agree a text.
  • Yemen, Sudan and Ethiopia — mass suffering with the Council captured by arms-supplying and patron states; and Gaza, where the charge of selectivity is now pressed most forcefully.
• The attempted repairs came from the doubters, not the opponents.
  • Brazil’s “Responsibility while Protecting” (2011) proposed that authorisations be sequenced, narrowly construed, time-limited and subject to monitoring and reporting back to the Council — a direct answer to Libya.
  • China’s “Responsible Protection” (2012) held that only the Council may authorise, that means must be limited to protection, that the intervener is accountable for consequences, and that regime change is never legitimate.
  • Neither has been adopted, and neither refuted; they mark the terms on which the doctrine could be rebuilt.
• The state of the debate in 2026. At the Assembly’s plenary debate on 6 July 2026, 50 member states and two regional organisations spoke for 89 countries, and the Group of Friends of R2P numbers 55 states.
  • 20 speakers representing 68 states criticised selective application, especially over Gaza. The argument is no longer whether the norm exists but whether it is applied honestly.

India’s Position

• India abstained on Resolution 1973 and remains sceptical of pillar three, while accepting pillars one and two without reservation. The position rests on four claims.
  • Authorisation must come from the Security Council and nowhere else — no coalition, regional organisation or doctrine of legitimacy may substitute for it.
  • Selectivity is the doctrine’s fatal defect in practice: intervention where the powerful have interests and inaction where they do not, which turns a universal norm into an instrument of the strong.
  • Regime change is not protection, and Libya showed that a protection mandate will be stretched to whatever the intervening coalition wants.
  • Capacity-building, not coercion, is the productive limb — India presses for investment in pillar two and in prevention rather than in the machinery for authorising force.
• It is rooted in the non-intervention tradition running from Panchsheel and non-alignment, in a post-colonial reading of intervention’s history, and in concern about precedents that could be turned toward internal conflicts.
• It is scepticism, not rejection. India accepts that sovereignty carries responsibility, votes for protection-of-civilians mandates and staffs them with more soldiers than anyone else; what it declines is that any state may decide unilaterally when a government has forfeited its sovereignty.

Why Collective Security Fails, and What Survives

The Consolidated Critique

The requirements set out earlier are not separate failures but one structural failure: the system asks states to treat an abstract obligation as more binding than a concrete interest, and gives enforcement to the states least likely to do so.

• The great powers exempted themselves, and the veto formalises the exemption — the system acts against the weak and never the strong, which makes it great-power policing with a universal membership attached.
• Identifying the aggressor is a political act. Who struck first is a question of fact only in the simplest cases; usually it is a question of where the sequence is said to begin, decided by whoever controls the vote.
• States subordinate the general obligation to the particular interest, and always have; this is the realist objection, and no example has answered it.
• The requirement to act against a friend has almost never been honoured. Every permanent member has vetoed to shield a client, and the pattern is invariant across ideology.
• The free-rider problem is structural. The guarantee is a public good, so each member gains whether or not it contributes and contribution collapses onto whichever states happen to care — visible now in the peacekeeping arrears crisis.
• The model assumes a clean case of interstate aggression that modern conflict rarely supplies. Civil wars, proxy wars, hybrid operations below the threshold of armed attack, cyber operations, unattributed irregulars, private military companies and non-state armed groups present no aggressor the system can name.
• Enforcement depends on the very powers most likely to be parties — a design fault rather than an accident of practice, since there is no version of the system in which the enforcers are disinterested.

The Honest Counter-Case

• The norm deters even where the machinery does not act: that aggression may be met by a coalition is a real cost in every calculation, and the wars it has priced out of consideration cannot be counted.
• Conquest has been delegitimised, a historically extraordinary achievement. Since 1945 forcible annexation has almost never been recognised — Kuwait was restored, Crimea’s annexation is rejected by most states — and the near-disappearance of conquest as a legitimate route to territorial change is collective security’s real legacy.
• Article 2(4) survives as a norm despite violation. That aggressors take elaborate trouble to call their wars self-defence, protection of nationals or invited assistance shows the prohibition constrains the argument even when it fails to constrain the act.
• The institutional residue is enormous — sanctions regimes, arms embargoes, war-crimes tribunals, peacekeeping, protection-of-civilians mandates, R2P and the ICC all sit in the space collective security opened and could not fill.
• It has succeeded at what it was not designed for — freezing conflicts, supervising settlements and giving weak states a forum in which they cannot simply be ignored.

The Veto and Reform, as They Bear on Collective Security

For the organisation’s structure, agencies and the general reform debate, see the dedicated article on the United Nations; what follows is confined to the veto’s effect on enforcement.

• The veto’s rationale was that the alternative was worse. The League gave every Council member a veto; the Charter narrows it to five. The great powers would not have joined an organisation able to coerce them, so the choice lay between a universal organisation with a veto and no universal organisation at all.
• It has a defensible modern form: a Council ordering enforcement against a nuclear-armed permanent member would be ignored, or would produce the great-power war the organisation exists to prevent.
• The veto switches the enforcement machinery off precisely when it is most needed. 2024 saw 7 vetoed drafts and 8 vetoes cast, the most since 1986 (Russia 4, the United States 3, China 1); 2025 saw 4 — two by the United States on Gaza, two by Russia on Ukraine.
• The restraint mechanisms are procedural rather than substantive, because nothing else is achievable without Charter amendment.
  • The Veto Initiative, General Assembly Resolution A/RES/76/262 of 26 April 2022, requires the Assembly President to convene a formal debate within 10 working days of any veto, obliging the vetoing state to explain itself to the whole membership. It imposes a reputational cost; it cannot reverse a veto.
  • The ACT Group’s Code of Conduct commits signatories to support timely action against genocide, war crimes and crimes against humanity; over 120 states have signed and none of the five.
  • The French-Mexican initiative proposes a voluntary undertaking by the permanent members not to use the veto in mass-atrocity situations, on a determination by the Secretary-General. It has broad backing and no support from the states whose restraint it seeks.
• The membership question runs in parallel.
  • The G4 — India, Brazil, Germany and Japan — seek permanent seats with expansion in both categories; Uniting for Consensus, led by Italy and Pakistan, opposes new permanent members and offers more elected seats; the Ezulwini Consensus demands two permanent African seats with the veto.
  • In July 2026 the Assembly again decided to continue the intergovernmental negotiations, with the African Union model tabled — after more than three decades there is still no negotiating text.
• The connection to collective security is direct. Enlargement without touching the veto changes who is in the room, not who can stop the machinery, and curtailing the veto needs the assent of the states it protects under Article 108. That is the closed circle in which every reform proposal ends.

Collective Security in 2026

• A permanent member is the aggressor in the largest interstate war in Europe since 1945, and the Council cannot even determine under Article 39 that aggression has occurred.
  • The war in Ukraine continues; a 32-hour Orthodox Easter truce in April 2026 collapsed amid hundreds of documented violations, with fighting resuming immediately, and there is no peace agreement. The system’s central case is the case it cannot reach.
• Gaza has produced an authorisation that looks less like collective security than great-power trusteeship. Resolution 2803 of 17 November 2025 was adopted 13-0, China and Russia abstaining, endorsing a Board of Peace and authorising an International Stabilization Force.
  • It is a Chapter VII authorisation, so the machinery moved — but it hands governance and security to an externally constituted board, displacing Palestinian self-determination and creating authority without accountability.
  • The pattern is now familiar: the Council authorises when the powerful have already settled the outcome elsewhere, and is silent when they have not.
• Peacekeeping, the substitute that kept the organisation relevant, is contracting under a funding collapse rather than because conflicts have ended.
• Regional organisations are taking over enforcement — the African Union in Somalia through AUSSOM, ECOWAS in West Africa despite the coup belt’s damage, and ad hoc coalitions elsewhere. Chapter VIII is becoming the main channel and Chapter VII the formality, decentralising enforcement to bodies whose impartiality is harder still to assert.
• The UN80 Initiative, launched by Guterres in March 2025, is the institutional response: workstreams on efficiencies, mandate review and delivery, General Assembly Resolution 80/251 of March 2026 on mandate creation and review, and a revised budget carrying a 21 per cent staff reduction. It consolidates; it does not reform enforcement.
• The concluding judgement. Eighty years on, collective security has never functioned as designed and has never been abandoned.
  • States keep rebuilding it because the alternative — pure self-help, in which conquest costs nothing beyond resistance — is worse for almost all of them, and because no state will publicly renounce a promise it privately declines to perform.
  • Its record is the record of an idea that organises the argument without ever winning it.

The Charter’s machinery has failed for eighty years and no state proposes to remove it. That tells us what collective security is now: not a guarantee, but the standard against which its own breach is measured.