Conflict
Why Groups Fight When Cooperation Would Win
Introduction
War is expensive. Not just in money and lives but in opportunity. Every resource spent on destruction is a resource not spent on building something. By almost any rational calculation, two nations that trade peacefully both end up richer than two nations that fight over the same territory. And yet wars keep happening. Conflicts erupt between groups that would clearly benefit from cooperation. Neighbors who share a border, an economy, and often a language and culture find themselves in cycles of violence that leave both sides worse off than where they started.
On the rationalist account associated with the political scientist James Fearon, this is not mainly because humans are inherently violent or irrational. The puzzle, on this account, is that conflict often emerges from rational actors making individually logical choices that produce collectively terrible outcomes. Understanding why requires looking at the conditions that make fighting seem preferable to negotiating: commitment problems, information failures, indivisible stakes, and the strange logic of credible threats. Conflict is then often reason operating under conditions that make peaceful agreement difficult to reach or sustain, rather than a breakdown of reason. Rival accounts stress psychology (misperception and overconfidence), domestic politics (leaders whose interests differ from their countries'), or shifts in the balance of power.
Game Theory of Conflict
Game theory provides a framework for understanding why rational actors sometimes choose destruction. Consider the classic prisoner's dilemma: two suspects are interrogated separately, and each has the option to cooperate (stay silent) or defect (betray the other). If both cooperate, both get a light sentence. If both defect, both get a harsh sentence. But if one defects while the other cooperates, the defector goes free while the cooperator gets the harshest sentence of all. The rational choice for each individual is to defect, regardless of what the other does, even though mutual cooperation produces a better outcome for both.
This structure appears in many real conflicts. Two countries in an arms race would both be better off reducing military spending and investing in their economies. But neither can trust the other to actually reduce, because if one side disarms while the other does not, the disarming side is vulnerable. So both keep spending, locked in a pattern that makes both worse off. Two ethnic groups sharing a territory might both prefer peaceful coexistence to civil war, but if each fears the other might strike first, both have incentives to arm and prepare, which looks threatening to the other side and confirms their fears. Rational caution produces the very outcome everyone wanted to avoid.
The key insight from game theory is that the problem is often not about intentions but about structure. Even groups with peaceful preferences can end up in conflict if the strategic situation makes cooperation risky and defection tempting. Changing outcomes requires changing the structure: making cooperation enforceable, making defection costly, or creating mechanisms for credible communication. Without structural change, appeals to goodwill and peace are not enough, because the underlying incentives still push toward conflict.
Escalation Dynamics
Conflicts rarely begin with a decision to go to war. They begin with a disagreement, a provocation, or a misunderstanding that escalates through a series of steps, each of which seems reasonable in isolation but leads collectively toward catastrophe. World War I is the textbook example: a political assassination triggered a chain of alliance obligations, mobilization schedules, and ultimatums that pulled an entire continent into a war that, on this reading, none of the major powers initially wanted. Each step made sense from the perspective of the actor taking it. The accumulated result was four years of slaughter. Historians still dispute this reading: Fritz Fischer's school sees the war as imperial Germany's bid for world power.
Escalation dynamics are driven by several reinforcing mechanisms. Commitment problems arise when neither side can credibly promise to stop escalating. If you back down after a provocation, the other side may learn that provocation works, which encourages more of it. So you respond, which creates a new provocation for them to respond to. Audience costs compound this; leaders who have publicly committed to a tough stance risk domestic political punishment for backing down, even when backing down would be strategically wise. Sunk cost thinking kicks in as the conflict progresses: each sacrifice already made creates pressure to continue rather than accept that those sacrifices were wasted. The logic of escalation creates its own momentum.
One of the most dangerous features of escalation is that both sides typically believe they are responding defensively. From each side's perspective, their actions are measured responses to the other side's provocations. But what looks defensive from inside feels offensive from outside. A military buildup you see as protection, your neighbor sees as preparation for attack. A retaliatory strike you consider proportional, the other side considers escalatory. This perceptual asymmetry, where both sides believe they are the reasonable one, is one of the most consistent features of conflicts throughout history and one of the hardest to overcome.
Resource Competition vs Ideological Conflict
Not all conflicts are driven by the same engine. Resource conflicts, over territory, water, oil, minerals, or arable land, operate on a logic of scarcity. If two groups need the same river for irrigation, and there is not enough water for both, they have a real zero-sum problem. These conflicts are conceptually simpler (though not easier) to resolve because they involve divisible stakes. You can split territory, share water rights, divide resource revenues. Negotiation is possible because both sides can do the math and find distributions that leave both better off than fighting.
Ideological and identity conflicts are different. When a conflict is about who controls a holy site, whose version of history is legitimate, or which group's identity defines a nation, the stakes feel indivisible. No proposal for dividing Jerusalem has yet satisfied both sides' claims to it. You cannot compromise on whether a genocide happened. Identity-based conflicts activate deep emotional commitments that make the calculating logic of bargaining feel obscene. Suggesting that both sides give up half of what they want sounds reasonable to outsiders but insulting to participants who see their claims as matters of justice, not negotiation. How fixed such stakes are is disputed: Stacie Goddard argues that indivisibility comes from how leaders justify their claims and is not inevitable, while Ron Hassner argues that sacred places are especially hard to divide.
In practice, most real-world conflicts involve both resource competition and identity dimensions, often intertwined. A conflict that begins over land or resources often develops an identity component as grievances accumulate and narratives of victimhood solidify. A conflict rooted in ethnic or religious identity often has material dimensions, such as control over jobs, government budgets, or natural resources, that provide the economic fuel for sustained mobilization. Understanding which engine is driving a particular conflict matters because the solutions are different. Resource conflicts need institutional arrangements for equitable distribution. Identity conflicts need recognition, narrative change, and sometimes simply time.
Civil Wars and Internal Conflict
Most twenty-first-century organized violence happens within countries, not between them. By the count of the Uppsala Conflict Data Program (UCDP), fighting within states caused more battle deaths than fighting between states in almost every year from 1989 to 2022. Russia's war in Ukraine has reversed that since 2023: it caused about 62% of the world's battle deaths in 2025, when conflicts between states reached eight, the most since records began in 1946. Outside Ukraine, most killing still came from conflicts within countries, above all Sudan's war, along with fighting in eastern Congo and Myanmar, insurgencies from the Sahel to Somalia, and cartel violence in Mexico. Yet most of what we say about war is shaped by inter-state-war thinking - armies, borders, generals, treaties - which fits these internal conflicts poorly.
Internal conflicts have their own dynamics. They tend to be longer than inter-state wars, frequently lasting a decade or more, because the parties cannot retreat to separate territories; they share the same country. They more often involve civilians as targets and as combatants, because the line between combatant and noncombatant is less defined. They are harder to end through negotiation because credible commitment problems are intense: rebels who lay down arms become vulnerable to a government that controls the security forces, and governments that grant amnesty risk creating incentives for the next group to take up arms. And they tend to be especially recursive: by one widely cited estimate, nearly half of all civil wars are relapses of earlier ones, often involving the same factions or fragments of them.
External actors complicate civil wars in distinctive ways. Foreign governments routinely arm, fund, and advise factions on whichever side aligns with their interests; in 22 of the 57 internal conflicts UCDP recorded in 2025, other governments also sent troops. This outside involvement (Russia in Syria, Saudi Arabia and Iran in Yemen, the United States in Afghanistan, regional powers across the Sahel) often extends conflicts beyond what the local parties alone could sustain, and tangles the negotiations needed to end them. The cost is borne overwhelmingly by people who live in the country and never asked for any of it.
The Security Dilemma
One of the most powerful concepts in conflict studies is the security dilemma: measures one state takes to increase its own security tend to decrease the security of its neighbors. You build a stronger military to protect yourself. Your neighbor, unable to read your intentions, sees a growing threat and builds up their own military in response. You see their buildup and conclude you were right to worry, so you increase your forces further. Both sides end up less secure and poorer, locked in a spiral that neither intended but neither can safely exit.
The security dilemma is particularly intense when offense and defense are hard to distinguish. If a weapon system that is devastating in attack is identical to one that is essential for defense, then any military buildup looks aggressive regardless of intent. Tanks massed near a border can be defensive preparation or invasion staging; there is no way to tell from observation alone. This ambiguity is why arms control agreements focus not just on quantities but on types and positioning of weapons, trying to create configurations that are clearly defensive and cannot easily be repurposed for attack.
The security dilemma also operates at smaller scales. In neighborhoods with high crime and low trust in police, individuals may arm themselves for protection. But each armed neighbor also looks like a threat to the others, which increases everyone's incentive to arm, which raises the danger further. Ethnic groups in a fragmenting state face the same logic: when the central government can no longer guarantee safety, each group must provide its own security, and the measures they take threaten other groups, triggering exactly the violence they feared. The tragedy of the security dilemma is that it can produce conflict between actors who want peace, because the structure of the situation makes peaceful intentions indistinguishable from hostile ones.
How Peace Is Maintained
Given all these pressures toward conflict, what keeps peace? Several mechanisms work, often in combination. Deterrence, making the cost of attack so high that no rational actor would attempt it, has been the dominant framework since nuclear weapons made great-power war potentially suicidal. The logic is straightforward: if attacking guarantees your own destruction, you will not attack. Nuclear deterrence has coincided with the longest period without direct war between major powers in modern history, though John Mueller has called the weapons essentially irrelevant to it. But deterrence can be fragile. It requires rational decision-makers, reliable communication, and functioning command-and-control systems. A single miscalculation, a false alarm, or an irrational leader could collapse the entire framework. The 2020s have strained this framework more than any decade since the 1980s. Russian leaders have publicly threatened nuclear use over Ukraine, and New START, which capped US and Russian strategic forces, lapsed in February 2026. North Korea has vowed to expand its arsenal "exponentially," and India and Pakistan, both nuclear-armed, fought a brief conflict in May 2025. Israel and the United States struck Iran in June 2025 and have been at war with it since February 2026, citing its nuclear program, which Iran says is peaceful. AI-assisted military decision-making and hypersonic weapons could also compress decision timelines below what humans can reliably manage. Nothing has broken yet. The assumption that nuclear deterrence is a stable equilibrium has become noticeably less safe.
Trade interdependence creates different incentives for peace. When two countries are deeply integrated economically, when factories in one country depend on components from the other, when banks hold each other's debt, when millions of jobs depend on continued commerce, war becomes economically catastrophic for both sides. This was the logic behind the European project: bind France and Germany together economically so tightly that war between them becomes unthinkable. It has worked remarkably well in Europe. The theory took a hit, though, when Russia's invasion of Ukraine demonstrated that deep trade relationships do not always prevent conflict, especially when political leaders prioritize other objectives over economic welfare. The US-China relationship is the live test case for what trade interdependence can and cannot do. In what is sometimes called "de-risking" rather than "decoupling," both sides are deliberately unwinding part of their economic integration for security reasons. Whether this deliberate weakening of interdependence will reduce the chance of conflict (by separating the two economies before a crisis) or increase it (by removing the mutual costs that have so far discouraged confrontation) is an open question being answered in real time.
International norms and institutions provide a third mechanism. The norm against territorial conquest, the idea that you cannot simply invade a neighbor and take their land, is relatively new in historical terms but has become remarkably strong. Dan Altman's data suggest a narrower change: attempts to seize parts of states stayed common after 1945, shifting to small land grabs meant to avoid war. Wars of conquest that were routine in the 1800s are now considered illegal under international law and trigger broad condemnation and sanctions. These norms do not prevent all aggression, but they raise the cost by ensuring that violators face diplomatic isolation, economic penalties, and reputational damage. No single mechanism guarantees peace. But layered together (deterrence, interdependence, norms, and institutions), they create a web of incentives that makes conflict more costly and cooperation more attractive. The challenge is maintaining all of these mechanisms simultaneously, because each one has vulnerabilities that the others need to cover.
Why Some Regions Stabilize While Others Don't
Western Europe went from the most war-torn region on Earth to one of the most peaceful in a single generation. Meanwhile, parts of the Middle East, Central Africa, and South Asia have experienced recurring conflict for decades. These divergent outcomes are less about culture or inherent tendencies toward violence than about structural conditions, though cultural factors are not irrelevant (as explored in the Law page's discussion of legal traditions). Europe stabilized through a specific combination of circumstances: American security guarantees that resolved the security dilemma, economic integration, through what became the EU, that created shared prosperity, democratic governance that gave populations voice in foreign policy, and the overwhelming deterrent of nuclear weapons. This combination was historically specific and not easily replicable. Some elements, particularly the expansion of Western security alliances, are themselves contested: stabilizing from one perspective, threat-generating from another.
Regions that remain volatile typically lack one or more of these stabilizing conditions. Borders drawn by colonial powers that do not correspond to ethnic, linguistic, or economic realities can create lasting internal tensions. Weak state institutions cannot credibly commit to protecting minority groups, triggering the security dilemma at domestic scale. Resource wealth can fund armed factions and create stakes worth fighting over. External powers intervene to support different sides, extending conflicts that might otherwise burn out. And once a cycle of violence begins, it generates grievances, displaced populations, and destroyed institutions that make the next round of violence more likely, a conflict trap that is extremely difficult to escape. Scholars weigh these causes differently: Paul Collier and Anke Hoeffler found that the feasibility of rebellion predicted civil war better than grievance did, James Fearon and David Laitin stressed weak states and rough terrain over ethnic diversity, and Lars-Erik Cederman and colleagues found that ethnic groups shut out of power rebel more often.
Understanding these factors matters because it shifts the conversation from blame to mechanism. Asking "why can't those people just get along" assumes the problem is cultural or moral. Asking "what structural conditions make peaceful cooperation rational and sustainable in this specific context" leads to actionable answers. Sometimes the answer involves economic development. Sometimes it involves security guarantees from external actors. Sometimes it involves institutional redesign that gives all groups a stake in the system. There is no universal template, but there are patterns that help explain why peace holds in some places and breaks down in others.
Terrorism as Asymmetric Strategy
On the strategic view that dominates terrorism studies, terrorism is not random violence but a choice by groups that are too weak to fight conventionally. A small militant organization cannot defeat a modern military on a battlefield. What it can do is provoke a disproportionate response. This is the core logic of terrorism as strategy: commit a dramatic act of violence that forces the stronger side to react in ways that generate sympathy for the attacker's cause, radicalize moderates, and impose costs that the stronger side eventually decides are not worth bearing.
This strategy works more often than people realize, not in achieving grand political objectives, which terrorist groups rarely attain, but in provoking the overreaction it seeks. When a government responds to terrorism with broad crackdowns, mass surveillance, or military operations that harm civilian populations, it often validates the terrorist group's narrative and drives recruitment. Political scientist Robert Pape studied every suicide attack from 1980 to 2001 and concluded that the campaigns aimed to coerce democracies into withdrawing forces from land the attackers saw as their homeland, suggesting strategic calculation more than irrational fanaticism. Critics reply that a sample of attacks alone cannot show their causes, that later tests found regime type unrelated to suicide terrorism, and that al-Qaeda's Salafi-jihadist ideology drove the attacks' spread in the 2000s. None of this makes terrorism morally acceptable. The point is that understanding it as calculation rather than madness leads to more effective responses.
The costs of terrorism extend far beyond direct casualties. Societies spend enormous resources on security measures, accept significant curtailments of civil liberties, and redirect political attention from other pressing issues. In the United States, the response to 9/11 cost trillions of dollars, launched two decades-long wars, and transformed domestic surveillance practices. The direct attack killed roughly 3,000 people. The response reshaped global politics. This disproportion between the act and the reaction is not a flaw in the strategy; it is the strategy. Understanding this does not mean accepting terrorism. It means recognizing that the most effective counter-terrorism approaches are often those that deny terrorist groups the overreaction they seek, rather than providing it.
Cyber and Gray-Zone Conflict
For most of human history, the line between war and peace was relatively clear: armies marched, blood was shed, treaties followed. Modern great-power conflict increasingly operates below that line, in what strategists call the gray zone, coercion that hurts an adversary without crossing the threshold that would trigger a conventional military response. Cyber attacks fit here. Disinformation campaigns fit here. Economic coercion, paramilitary deniability, sabotage of undersea cables, election interference, intelligence operations against foreign critical infrastructure, all of it lives in the gray zone.
Cyber capability has become a standard tool of state behavior. Nation-states routinely probe each other's networks, embed access points in foreign critical infrastructure for potential later use, and conduct espionage at scales unimaginable a generation ago. Some operations are visible: the 2014 Sony hack attributed to North Korea, the 2017 NotPetya attack attributed to Russia, recurring intellectual-property theft attributed to China, and the Stuxnet sabotage of Iranian nuclear centrifuges, reported to be a joint US-Israeli operation. Most are not. Attribution is technically difficult, often deniable, and politically costly, which is exactly why this domain is attractive: a state can do real damage without ever being conclusively linked to it.
Gray-zone conflict creates a problem for traditional deterrence frameworks. Deterrence assumes the adversary believes that crossing a particular line will trigger a particular response. Gray-zone activity is designed to never cross any single line large enough to trigger a response, while collectively imposing significant costs over time. Russia's 2014 operations in Crimea and eastern Ukraine, using uniformed soldiers without insignia, paramilitary groups, and information operations alongside formal military movement, were a textbook case. The Western response was significant but slow, in part because the operation never produced a single clear moment that demanded escalation. Whether democracies can develop adequate responses to this kind of pressure, without overreacting in ways that play into adversaries' hands, is one of the open security questions of the next decade.
What Works
Ending conflicts is harder than starting them. Negotiations often fail, ceasefires collapse, and peace agreements unravel. Research on conflict resolution identifies several factors that make success more likely. First, what I. William Zartman calls a mutually hurting stalemate, a situation where both sides are suffering and neither can win, creates the conditions for serious negotiation. As long as one side believes it can prevail through continued fighting, it has little incentive to make the concessions peace requires. Bargaining theorists add that fighting reveals which side is stronger, so wars tend to end once the two sides' expectations converge. This is why well-intentioned calls for negotiation often fail early in a conflict: the parties have not yet reached the point where peace looks more attractive than continued fighting.
Second, credible commitment mechanisms are essential. The fundamental problem in ending a conflict is that both sides must disarm or demobilize, but whoever disarms first is vulnerable. Peace agreements that include external security guarantees (peacekeeping forces, monitoring missions, third-party enforcement) are significantly more likely to hold than agreements that rely purely on the parties' promises to each other. Monica Toft finds, however, that civil wars ended by negotiated settlement have recurred more often than those ended by victory, especially rebel victory, though others dispute her coding. Power-sharing arrangements that give former combatants a stake in the new political system reduce the incentive to return to fighting. Transitional justice mechanisms (truth commissions, amnesties, tribunals) address grievances that could otherwise reignite conflict, though scholars debate which approach works best in which context.
Third, and most difficult, lasting peace requires addressing the underlying structural conditions that produced the conflict. If a war was driven by exclusion of an ethnic group from political power, peace requires institutional reform that guarantees inclusion. If it was driven by competition over resources, peace requires equitable distribution mechanisms. If it was driven by security fears, peace requires credible arrangements that make both sides feel safe. These changes take years or decades, which is why post-conflict societies stay fragile for a long time. Peace is not a moment. It is a process, and it requires sustained investment in the conditions that make cooperation more rational than fighting.
A Note at the End
If you have read through this site, or even just parts of it, you now carry a clearer picture of how these systems work than most people typically encounter. You understand that your brain constructs reality rather than recording it. That decisions are shaped by invisible biases. That trust is fragile and institutions are imperfect. That markets, money, and power operate on logic that rarely matches the simple stories people tell about them. That conflicts arise from structure, not just malice. None of this knowledge is comfortable. Understanding how systems actually work can feel heavy, because you start seeing the machinery behind things you used to take at face value.
But here is something worth remembering: most of what happens in the world is beyond your control. You did not design these systems. You cannot single-handedly fix them. And spending your energy being outraged at every injustice, every manipulation, every broken institution is a fast path to exhaustion. Not because these things do not matter, but because anger without action is just wear and tear on your nervous system. Understanding the world clearly is valuable. Letting that understanding consume you is not.
What you can control is yourself. Take care of your body. It is the only hardware you get, and it runs better with movement, sleep, and reasonable fuel. Train your mind to evaluate events with calm analysis rather than reactive emotion. Not because emotions are bad (they are essential, as the earlier pages explained) but because the ability to pause between stimulus and response is one of the most useful skills a human can develop. When you see a headline designed to make you angry, notice the design. When someone tries to manipulate you, recognize the technique. When a system seems broken, understand the incentives that keep it that way. And then decide, with a clear head, what is actually worth your energy. Not everything is. Choose well.
