Iran and Pakistan exchanged military strikes across their shared border this week in one of the most serious confrontations between the two neighbours in years. On Tuesday, Iran launched missiles and drones into Pakistan’s southwestern Balochistan province, saying it had targeted bases belonging to Jaish al-Adl, a Sunni militant organisation responsible for attacks against Iranian security forces. Pakistan said the strike violated its sovereignty and killed two children. Islamabad recalled its ambassador and, two days later, launched its own strikes inside Iran, saying it had targeted hideouts belonging to Baloch separatist organisations threatening Pakistan. The extraordinary feature of the confrontation is that Iran and Pakistan were not attacking each other because they fundamentally disagree about whether the militants should be suppressed. Both governments oppose separatist and militant organisations operating around the border. They attacked each other because neither trusted the other to suppress those organisations effectively enough.
That makes this week’s confrontation a useful test of what sovereignty actually means when governments do not exercise equal control over every part of their territory. On a political map, the Iran-Pakistan border is perfectly clear. Pakistan possesses sovereignty westwards until the border and Iran possesses sovereignty eastwards from it. International law therefore begins with a straightforward principle: Iran cannot simply launch missiles into Pakistan because it believes a target there threatens Iranian security, and Pakistan cannot do the reverse. Borders would become meaningless if every government could decide independently when another state’s enforcement efforts were insufficient and then substitute its own military force.
The practical problem is that armed organisations do not necessarily respect those borders. Jaish al-Adl operates around Iran’s southeastern frontier and has carried out attacks against Iranian security forces. Baloch separatist organisations fighting Pakistan operate across a region whose ethnic and geographical boundaries do not correspond neatly with the international border. The Baloch population spans both Pakistan’s Balochistan province and Iran’s Sistan-Baluchestan province, while the border passes through a large, relatively underdeveloped and difficult-to-govern region. Militants can therefore exploit precisely the difference between legal borders, which are clear, and administrative control, which is much less complete.
This creates a distinction between formal sovereignty and effective sovereignty. Formal sovereignty is binary: territory internationally belongs to one state or another. Effective sovereignty is variable: governments exercise different levels of administrative, security and economic control across different regions. Islamabad possesses the same legal sovereignty over remote parts of Balochistan as it possesses over central Islamabad, but the practical capacity of the state to monitor every kilometre of a sparsely populated frontier is obviously different. Iran faces a similar challenge in Sistan-Baluchestan. Militants survive inside the gap between the sovereignty states claim and the capacity they actually possess.
The international system nevertheless needs formal sovereignty to remain the default rule because the alternative would be extraordinarily destabilising. If a government could attack whenever it concluded that another state was insufficiently capable of policing its territory, powerful countries would effectively acquire greater rights than weak ones. States with sophisticated militaries could decide repeatedly that fragile neighbours had failed their security obligations and conduct operations inside them. Sovereignty would cease functioning as protection against external power precisely for countries that need that protection most.
This is why Iran’s justification for striking Jaish al-Adl cannot by itself resolve the legal or political problem. Tehran may possess genuine evidence that militants operating from Pakistan have attacked Iranian personnel. That gives Iran a legitimate security grievance. It does not automatically give Iran jurisdiction over Pakistani territory. The normal response should be intelligence sharing, border cooperation, extradition where appropriate and pressure on Pakistani authorities to act. Unilateral force should remain exceptional precisely because once the precedent becomes normal, other governments can invoke the same argument against Iran.
Pakistan demonstrated that immediately. Islamabad’s retaliation used essentially the same logic Iran had used two days earlier: militants threatening Pakistan were allegedly operating inside Iranian territory, therefore Pakistan struck them directly. Each government could therefore describe its own attack as counterterrorism and the other’s as a violation of sovereignty.
That symmetry reveals the weakness of unilateral doctrine.
If the principle is “we may strike militants abroad when the host state does not remove them”, it cannot logically belong to only one government.
Iran cannot claim the principle on Tuesday and reject Pakistan’s use of it on Thursday.
Pakistan cannot condemn Iran’s violation while pretending its retaliation raises no equivalent sovereignty question.
Rules become rules only when governments accept them while they constrain their own actions.
This is another veil-of-ignorance test. A government deciding what international rule should apply needs to imagine itself as both the state pursuing militants and the state whose territory contains them. Rules designed only from the first perspective will authorise intervention too easily. Rules designed only from the second could allow governments to ignore armed organisations using their territory to attack neighbours. A stable international system therefore needs both territorial sovereignty and obligations of due diligence, meaning states should make genuine efforts to prevent their territory from becoming a base for attacks on others.
The difficult part is determining what happens when those obligations fail.
International law and diplomacy provide several mechanisms short of unilateral military force, but those mechanisms can feel inadequate when soldiers or civilians continue dying. Governments then face domestic pressure to demonstrate that they can protect their populations regardless of whether neighbouring states cooperate.
Iran experienced exactly that pressure after attacks by militant organisations, including the killing of 11 Iranian security personnel in December in Sistan-Baluchestan. From Tehran’s perspective, another diplomatic complaint can appear meaningless if the organisation responsible remains capable of operating.
This is where state credibility and international restraint can come into conflict. A government needs citizens and adversaries to believe attacks will produce consequences. Yet the most immediate method of demonstrating resolve, striking the suspected sanctuary directly, can violate another state’s sovereignty and create a much larger security problem.
Counterterrorism therefore contains its own escalation dilemma.
The immediate target may be a non-state organisation.
The territory belongs to another state.
Once that state retaliates, the conflict moves from government-versus-militant towards government-versus-government.
The stronger the original counterterrorism operation, the greater the possibility of producing an interstate confrontation much more dangerous than the militant threat it was supposed to address.
This is precisely what happened this week. Iran’s strike was directed, according to Tehran, against Jaish al-Adl rather than the Pakistani government. Pakistan nevertheless could not treat missiles entering its territory as though Pakistani sovereignty were irrelevant. If Islamabad had simply accepted the strike, future Iranian operations could become easier politically and other adversaries might infer that Pakistan tolerates external violations under pressure.
Pakistan therefore faced a reputational incentive to retaliate.
That incentive was particularly strong because Pakistan is a nuclear-armed state with a powerful military whose domestic legitimacy depends substantially on presenting itself as the defender of national sovereignty. Allowing a neighbouring state to conduct an unprecedented missile strike without response could have imposed substantial political costs on the military leadership.
Retaliation therefore served several audiences simultaneously.
It attempted to deter Iran.
It reassured Pakistani citizens.
It demonstrated military capability.
It signalled to other states that Pakistani territory could not be attacked without consequence.
This is why military decisions cannot be understood solely through their effect on the immediate adversary.
Governments signal to multiple audiences at once.
The danger is that Iran faces the same reputational incentives.
If Pakistan retaliates and Iran does nothing, Iranian leaders could fear appearing weak after initiating the confrontation.
This is how limited actions acquire escalation momentum.
Each state believes it needs one additional action to restore the deterrence balance disturbed by the previous one.
The confrontation therefore resembles the Middle Eastern escalation problem from the last two weeks, but with an important difference: Iran and Pakistan appear to have very little strategic reason to fight each other directly. They maintain diplomatic and commercial relations, neither has a territorial claim requiring war, and both face more important security problems elsewhere. Pakistan’s military remains heavily focused on India and domestic militancy. Iran is deeply involved in confrontation surrounding Israel, the United States and its regional network of partners.
A major Iran-Pakistan conflict would therefore divert resources from both countries’ more important priorities.
That creates strong incentives for de-escalation.
Those incentives appear to be operating. Pakistani and Iranian officials have already begun signalling that neither side seeks a broader conflict, while diplomatic contacts continue. This is important because de-escalation also requires signalling. Governments cannot simply stop firing if they fear silence will be interpreted as weakness. They need narratives allowing both sides to claim that deterrence has been restored and therefore additional retaliation is unnecessary.
This is what diplomatic off-ramps actually do.
They do not always resolve the underlying dispute.
They create a politically acceptable explanation for stopping.
Iran can say it demonstrated that Jaish al-Adl cannot attack without consequence.
Pakistan can say it demonstrated that violations of Pakistani sovereignty will receive a response.
Both can then return to diplomacy without either formally admitting defeat.
That may sound like face-saving theatre.
Face-saving is strategically valuable when the alternative is another missile strike.
International relations frequently depends on allowing governments to de-escalate without humiliating themselves.
Humiliation increases the domestic cost of restraint.
Diplomacy lowers it.
The deeper problem, however, remains untouched. Militants still operate around the border because the border region contains structural conditions that missile strikes cannot solve. Balochistan on the Pakistani side and Sistan-Baluchestan on the Iranian side are both peripheral regions with significant poverty, underdevelopment and longstanding grievances towards central governments. The specific militant organisations differ in ideology and objectives, but both states confront forms of armed mobilisation emerging from regions where political and economic integration into the centre remains incomplete.
This does not mean poverty automatically causes terrorism or separatism. Most poor people do not join armed groups, and ideological, ethnic, religious and geopolitical factors matter enormously. The relevant mechanism is that weak development and limited state presence can reduce the opportunity cost of militancy while making armed organisations harder to monitor.
A young person with stable employment, access to public services and confidence in political institutions faces different incentives from one living in a region where the state appears principally through security forces.
Development therefore becomes part of counterinsurgency.
Roads matter.
Schools matter.
Employment matters.
Courts matter.
Political representation matters.
None replaces security operations against organisations conducting violence.
They affect the environment determining whether those organisations can recruit and regenerate after security operations occur.
This is why missiles can destroy militant infrastructure without solving militancy.
A training site can be rebuilt.
A commander can be replaced.
The underlying political market for armed mobilisation may remain.
Governments therefore need to distinguish degrading an organisation from changing the equilibrium that sustains it.
The United States learned versions of this problem across Afghanistan and Iraq. Enormous military superiority allowed repeated elimination of militant leaders and infrastructure. Political organisations adapted because military capability was only one input into their survival.
Iran and Pakistan face the same logic at a smaller regional scale.
Cross-border cooperation would therefore probably produce greater long-term security than reciprocal strikes. Intelligence sharing can make sanctuaries harder to use. Coordinated patrols can reduce border mobility. Economic investment can increase the state’s presence through institutions other than the military.
But cooperation requires trust.
This week’s attacks have reduced exactly that resource.
That creates another paradox.
The less governments trust one another to suppress militants, the more likely they become to act unilaterally.
Unilateral action violates sovereignty and reduces trust further.
Lower trust makes future unilateral action more likely.
A security problem therefore produces behaviour that worsens the institutional conditions required to solve it.
Breaking the cycle requires credible reciprocal commitments. Pakistan needs confidence that Iran will act against organisations threatening Pakistan from Iranian territory. Iran needs equivalent confidence regarding Jaish al-Adl. Verification matters because neither government will rely indefinitely on assurances from the other.
This is an area where practical security institutions can accomplish more than dramatic diplomatic declarations.
Joint border mechanisms.
Hotlines.
Intelligence exchanges.
Procedures for notifying operations.
Regular military contacts.
These arrangements appear mundane compared with missile strikes.
That is why they are valuable.
Institutions reduce the number of situations requiring leaders to make high-stakes decisions under uncertainty.
The confrontation also illustrates why the wider Middle Eastern crisis is becoming dangerous even when individual events are not directly caused by Gaza. Iran’s strike on Pakistan followed Iranian missile attacks in Iraq and Syria, while Tehran was responding partly to its own recent security shocks, including the Islamic State bombing in Kerman earlier this month. The region therefore contains several overlapping conflicts operating simultaneously: Israel-Hamas, Israel-Hezbollah, Red Sea attacks, US-Iran tensions, Islamic State violence, Kurdish conflicts and Baloch militancy.
These conflicts are not one war.
They interact.
That distinction matters because describing everything as one regional conflict exaggerates central coordination, while treating every event separately misses how heightened insecurity changes risk tolerance across the entire system.
Iran currently faces pressure from several directions.
That may increase its willingness to demonstrate military reach because leaders want adversaries to believe the state remains capable of retaliation.
Pakistan then becomes affected by a regional security environment it did not create.
This is how geopolitical instability spreads.
Not every transmission occurs through alliances.
Some occurs through changes in strategic behaviour produced by fear elsewhere.
The Iran-Pakistan exchange therefore offers a useful counterpoint to the Houthi confrontation. In Yemen, the United States is attempting to deter a non-state organisation through military punishment. Along the Iran-Pakistan border, two states are each attempting to deter non-state organisations while simultaneously deterring each other from violating sovereignty.
Deterrence becomes layered.
Pakistan wants Baloch militants deterred.
Iran wants Jaish al-Adl deterred.
Pakistan wants Iran deterred from crossing the border.
Iran wants Pakistan to prevent militants from operating there.
Every actor is trying to change another actor’s behaviour.
The resulting system can become less secure despite everyone claiming defensive objectives.
This is the security dilemma multiplied across state and non-state actors.
It also demonstrates why territorial sovereignty remains valuable even when it feels inefficient. Borders force governments to cooperate because neither can simply exercise police powers everywhere it believes threats exist.
That constraint can be frustrating.
It prevents counterterrorism from becoming a permanent licence for stronger states to operate militarily inside weaker neighbours.
The alternative would produce a world where every insurgent sanctuary becomes a potential interstate battlefield.
There are circumstances in which states argue that self-defence permits force against non-state actors operating from another country’s territory, particularly when the territorial state is unwilling or unable to address the threat. The legal status and scope of that doctrine remain contested internationally.
The political danger is easier to see.
Every government prefers a doctrine giving itself broad freedom to pursue enemies abroad and narrow freedom to everybody else.
Iran and Pakistan have just demonstrated what happens when both sides apply the broader version simultaneously.
The result is not more sovereignty.
It is reciprocal vulnerability.
This is why the most important development from this week may ultimately be whether both governments stop after one exchange.
If they do, the strikes will still have damaged relations and raised serious legal questions, but they may also reveal that limited retaliation can sometimes restore a deterrence equilibrium without escalating indefinitely.
That outcome is not guaranteed.
It depends on leaders recognising when enough signalling has occurred.
Strategic restraint often means stopping before the opponent has been completely punished.
That can feel unsatisfying domestically.
It is frequently what prevents limited conflict becoming war.
Pakistan does not need to defeat Iran.
Iran does not need to defeat Pakistan.
Both need militants operating around their border to become less capable of attacking them.
Once that objective is remembered, bombing one another begins looking like an extraordinarily inefficient substitute for cooperation.
The border itself is not the problem.
The gap between the sovereignty drawn along it and the state capacity exercised around it is.
Iran attempted to solve that gap with missiles.
Pakistan answered with missiles of its own.
Neither strike created more effective governance in Balochistan or Sistan-Baluchestan.
Neither created jobs.
Neither strengthened local political legitimacy.
Neither made the border easier to administer.
They demonstrated resolve.
Resolve is sometimes necessary.
It is not state capacity.
The long-term solution requires both countries to make the territory on their own side of the border harder for armed organisations to exploit, while trusting the other enough that security cooperation becomes more attractive than unilateral force.
That is slower than firing missiles.
It is also much more likely to solve the problem the missiles were supposed to address.
