ROGER PARTRIDGE: When Intervention Is Justified … or Not

“If not military intervention, then what? And when is intervention justified?” Those were the challenges from readers of my recent essay arguing conservatives should not be too quick to praise President Trump’s removal of Venezuelan dictator Nicolás Maduro.

My objection was not that Maduro did not deserve his fate – he did. It was that methods matter. The Venezuela operation was unilateral, without congressional authorisation or allied support. Its justification was openly transactional – oil, drugs, the Monroe Doctrine. There was no plan for what follows.

The historian Niall Ferguson, writing in The Free Press, praised Trump as a “nineteenth-century figure” returning to the politics of 1900. But that is not a compliment. The politics of 1900 produced 1914, and then 1945. A rerun with nuclear weapons will be worse.

It is easy to say, “not like this.” It is harder to say when intervention is justified – and why. Trump has made the question urgent. Venezuela was not a one-off. Trump has already warned Cuba to “make a deal before it’s too late.” Colombia has been threatened over drug policy. The “Donroe Doctrine,” as Trump styles it, is the announcement that the Western Hemisphere is America’s sphere of influence, where Washington decides which governments are acceptable and removes those that are not.

This essay attempts an answer. Not a complete theory of humanitarian intervention, but a framework for thinking about the question. When can military force be justified? What criteria should it meet? And why should conservatives, of all people, care about constraints on power even when the target is an authoritarian monster?

Why Intervene at All?

Implicit in the question of when to intervene is the question of why. The narrow answer is: only when it serves the national interest. Dwight Eisenhower gave a better one when he argued for American leadership of NATO in 1951. He called it “enlightened self-interest” – the recognition that American security depended on European stability, and European stability depended on American commitment. The costs of engagement were real. The costs of withdrawal would be greater.

Enlightened self-interest explains much. It explains why Western states intervened in Kosovo – instability in the Balkans threatened European security. And it may help explain why they failed to act in Rwanda, where the strategic stakes seemed lower.

Atrocity shocks the conscience even when no national interest is engaged – but not, it seems, enough to compel action when the victims are distant and the risks are high.

The concentric circles of moral obligation do not stop at the border. They thin as they extend outward – from family to community to nation to allies to humanity – but they do not vanish entirely.

The case for intervention rests on both pillars: the defence of an international order that serves the national security of liberal democracies, and the recognition that human suffering matters – particularly on a mass scale, and especially when it occurs in regions with which liberal democracies are deeply entangled by history or values.

Enlightened self-interest is not naïveté about costs. Modern history offers ample evidence that even well-intentioned interventions can harden resistance, prolong conflict, and impose burdens that far outlast the original justification. Vietnam, Afghanistan, Iraq and Libya differ in context and cause, but they share a lesson: once force is used, control over outcomes is quickly lost. The risks of acting are not hypothetical, and they fall not only on those intervened against, but on the interveners themselves. That is precisely why prudential thresholds must be high – and why frameworks matter.

The Landscape of the Right

Three positions dominate conservative thinking on military intervention.

The first is realism. John Mearsheimer and his disciples view humanitarian justifications as window dressing for great-power competition. States pursue interests, not principles. The rules-based order was always a fiction maintained by American hegemony, and now that hegemony is fraying.

This diagnosis has force. But realism offers no criteria for humanitarian intervention – because it does not believe humanitarian motives are real. States act from interest. The rest is pretence.

The second position is neoconservative interventionism. In 1996, William Kristol and Robert Kagan called for a “neo-Reaganite foreign policy” based on “benevolent global hegemony.” American power was good for America and good for the world. The United States should use that power confidently to promote democracy and confront hostile regimes. This vision animated much of the Bush administration’s foreign policy, culminating in Iraq. The lesson of Iraq was not that intervention is always wrong, but that enthusiasm without constraint is dangerous. Kristol and Kagan offered no limiting principles. Their framework could justify almost anything.

The third position is what we might call Trumpian transactionalism. This dispenses with both realist caution and neoconservative idealism. Trump does not pretend to be spreading democracy. He does not invoke humanitarian principles. He simply asserts American power openly and without apology. Ferguson finds this “refreshingly honest” compared to the hypocrisy of liberal internationalism. But honesty about abandoning principles is not a virtue. “We intervene when we want to” is not a framework. It is the absence of one.

What is missing from this landscape is a classical liberal account – one that takes seriously both the case for intervention and the case for constraint. Classical liberals are not pacifists. They recognise that force is sometimes necessary to defend the conditions under which free societies can flourish. But they also understand that unconstrained power is dangerous. That is the entire case for limited government, separation of powers, and the rule of law. The same logic applies to the international order.

A Framework for Justified Intervention

The question of when military intervention is justified is not new. The just war tradition runs from Thomas Aquinas through to Michael Walzer. In 2001, the International Commission on Intervention and State Sovereignty developed what became known as the Responsibility to Protect. R2P was adopted by the UN General Assembly in 2005 and remains part of the international vocabulary.

R2P sets out six criteria: just cause (large-scale loss of life or ethnic cleansing), right authority (Security Council authorisation or, failing that, General Assembly or regional bodies), right intention (the primary purpose must be to halt suffering), last resort (diplomatic options must be exhausted), proportional means (the scale of intervention must match the objective), and reasonable prospects of success (intervention should not make things worse).

These criteria represent the distilled wisdom of centuries of thinking about the morality of war. The problem is not the principles, but the machinery designed to apply them.

The first problem is institutional. R2P channels intervention through the UN Security Council, which is now paralysed by great-power rivalry. Russia and China will veto anything that resembles Western regime change after the Libya intervention in 2011, which they saw as a bait-and-switch – authorised to protect civilians, used to topple Gaddafi. The Security Council route is blocked and will not be unblocked any time soon.

The second problem is political. R2P has been championed primarily by liberal internationalists, wrapped in UN-centric language that alienates conservatives and leaves the centre-right without a framework of its own. If the only vocabulary for justified intervention is the vocabulary of the UN, then those who distrust the UN have no vocabulary at all – and default to either paralysis or raw assertion of power.

What follows is not a rejection of R2P but an adaptation of it – a framework for a world in which the Security Council cannot function as gatekeeper. It keeps what R2P gets right and adjusts what no longer works.

Just cause remains essential. There must be an ongoing atrocity, not merely bad governance. Genocide, ethnic cleansing, mass atrocities in real time. The intervention aims to stop killing that is happening, not to punish past misrule or secure resources.

This is a high threshold – and should be. Plenty of regimes are brutal, corrupt and authoritarian. That does not make them legitimate targets for military action. The bar must be set at the most extreme violations of human dignity. There are slow-motion catastrophes – like Venezuela’s collapse – that inflict enormous suffering but do not meet this threshold. That is precisely why they are not obvious candidates for military intervention.

Right authority needs revision. If the Security Council is blocked, the alternative is not unilateral action but collective action among democracies, arrived at through deliberation. Nineteen NATO members debating for months is different from one president acting on impulse. Collective action constrains adventurism. It forces leaders to justify their plans to sceptical allies. It builds legitimacy and shares burdens.

The authority comes not from a UN resolution but from the consent of free states acting together. The same logic applies domestically: a president who bypasses Congress on war is doing abroad what those who care about limited government would never tolerate at home – concentrating power in the executive and escaping the discipline of deliberation.

Right intention asks whether the intervener’s motives are genuinely humanitarian. That matters, but motives are hard to verify. The real work is done by the other criteria – particularly just cause and manifest loss of legitimacy. If those are met, mixed motives are tolerable. If they are not, pure motives cannot save the intervention.

Last resort and proportional means remain sound and need no revision.

Reasonable prospects of success should be strengthened into a requirement for a credible plan for what follows. International administration, supervised transition, regional integration as a goal. Regime change as part of a political settlement, not regime change as spectacle. The hardest part of any intervention is not the military operation but the aftermath. Without a plan, you create a vacuum that is often filled by forces worse than what you removed. R2P gestures at this with “reasonable prospects,” but the historical record – Libya, Iraq, Afghanistan – suggests the criterion needs teeth.

History suggests a further criterion not explicit in R2P should be added: manifest loss of legitimacy. There should be strong evidence that the regime has lost the consent of its people – stolen elections, mass population flight, systematic repression. This matters because it shapes the prospects for a stable aftermath and distinguishes intervention from conquest. A regime that retains genuine popular support, however distasteful its policies, is a poor candidate for intervention. A regime that rules only by terror and fraud is a different case.

These criteria are an appeal to prudence – to the recognition that even justified ends can be undermined by unjustified means.

What About Everything Short of Force?

None of this implies that democracies are powerless short of military intervention. The tools available – targeted sanctions, Magnitsky laws, financial restrictions, asylum for dissidents, diplomatic isolation, support for civil society, including secure communications – are not nothing. They are slow. But the comparison should not be with the satisfying speed of a raid. It should be with the decades of patient pressure that eventually brought down the Soviet bloc. That took forty years.

The rules-based order was never designed to topple dictators on demand. It was designed to prevent great-power war and to give smaller states space to develop without being devoured. Regime change was always the hardest case. The honest liberal position may be that some dictators fall only when internal conditions change – and that outside intervention often makes those conditions worse rather than better.

Libya is the cautionary tale.

None of this is as satisfying as watching a dictator hauled off in handcuffs. But satisfaction is not the test.

Two Cases

How does the framework apply in practice? Consider two cases: Kosovo in 1999 and Venezuela in 2026.

NATO’s intervention in Kosovo lacked UN Security Council authorisation. Russia would have vetoed any resolution. Critics at the time called it illegal. But examine it against the criteria.

There was an ongoing atrocity. Slobodan Milošević’s forces were engaged in ethnic cleansing of Kosovar Albanians. Refugees were streaming across borders. Mass graves were being dug. The intervention aimed to stop killing that was happening. There was multilateral consensus. Nineteen NATO democracies reached a consensus before acting. The decision was collective, not unilateral. Allies shared the burden and the risk. The justification was principled. NATO explicitly invoked humanitarian grounds. The intervention was framed as protection of civilians, not acquisition of territory or resources.

Diplomatic options had been exhausted. The Rambouillet negotiations collapsed when Milošević refused terms that every other party had accepted. Last resort was satisfied – not as a mere formality, but as genuine evidence that force had become the only remaining option. The means were proportional. NATO conducted an air campaign calibrated to degrade Serbian military capacity without a full-scale invasion. The force used matched the objective of halting ethnic cleansing, not conquering territory.

There was a plan for what followed. Kosovo came under UN administration, then supervised independence, with European integration as the long-term goal. The aftermath was imperfect, but there was a theory of the case. And the regime had manifestly lost legitimacy. Milošević’s brutality was documented and undeniable.

Kosovo met the criteria. Most in the West supported the intervention despite its lack of UN authorisation.

Now consider Venezuela. One president acted alone, without congressional authorisation, without allied support, without international mandate. The humanitarian situation was dire – economic collapse, mass emigration, stolen elections – but it fell short of genocide or mass slaughter. By design, the bar here is high. Trump’s justification was openly transactional: oil, drugs, the Monroe Doctrine. Diplomacy was abandoned, not exhausted. There was no plan for what followed beyond Maduro’s removal. The operation was targeted, not a full-scale invasion. But even granting proportional means alongside Maduro’s manifest loss of legitimacy, that is still only two criteria out of seven. The operation fails on the other five.

The contrast is not an embarrassing inconsistency. It is the framework working as intended.

Why Constraints Matter

The realist will object: constraints are for the weak. If America has the power to remove a dictator, why not use it? Who cares about frameworks and criteria when the result is one fewer tyrant?

Two answers.

First, the symmetry problem which I explored in that earlier essay. If America claims a sphere of influence in the Western Hemisphere, where it decides which governments are acceptable and removes those that are not, it validates identical claims by other great powers. A precedent of seizing oil and toppling a neighbour in the name of a nineteenth-century doctrine will not be read as an exception; it will be read as a model. Russia gets Eastern Europe. China gets the South China Sea and Taiwan. The rules-based order was built precisely to prevent this – to create a system in which even powerful states are bound by constraints, and smaller states have room to breathe.

Second, and more fundamental: those who believe in limited government should understand better than anyone that unconstrained power is dangerous. That is the entire case for constitutionalism, for separation of powers, for the rule of law. The point is not to trust concentrated power even when it is wielded by allies, because power corrupts and precedents outlast the individuals who set them.

The same logic applies internationally. The constraints that liberals built after 1945 – the UN Charter, the prohibition on aggressive war, the architecture of alliances and treaties – were not sentimental gestures. They were hard-won lessons from two world wars. They reflected the recognition that a world in which great powers do whatever they can get away with is a world that tends toward catastrophe.

Trump is now dismantling those constraints. His defenders say the old order was hypocritical, inconsistently applied, often ignored. They are right. But the answer to imperfect rules is not no rules at all. That is like arguing laws against murder do not work because murders still occur. The question is not whether the rules were violated. Of course they were. The question is what the world would have looked like without them. The post-1945 order produced a ninety per cent reduction in deaths from interstate conflict. Democracies expanded. Poverty declined faster than in any previous era. Those gains were not automatic. They rested on the principle that rules matter — and on the willingness to enforce them.

The framework also matters because Venezuela will not be the last case. As events in Iran already suggest, the question will not be whether regimes are loathsome, but whether the conditions for justified intervention are genuinely met.

Maduro deserved his fate. But “he deserved it” is not a framework. This essay has tried to offer one – criteria for when force is justified, and an honest account of what democracies can do when it is not.

This column was first published in Quadrant magazine on 23 January 2026.

Roger Partridge writes at Plain Thinking.