Peace negotiation

Why can two sides prefer a costly conflict even when both would benefit from peace?

The easy explanation for war is that somebody was irrational, or wicked, or badly informed. Hold that aside and notice how strange the phenomenon becomes if we assume the opposite — that both sides are competent, well advised, and want the best for themselves. Fighting destroys the thing being fought over, kills the people who would enjoy it, and ends, nearly always, in a negotiated settlement. So why not go straight to the settlement and keep the corpses and the treasure?

The puzzle is not meant to excuse anyone. It is that if war between capable adversaries follows from ordinary strategic logic, then "they were irrational" is a weaker explanation than it sounds — and a far less useful one, because you cannot design institutions around it.

Set up the accounting. Two sides dispute something divisible, and each has some chance of winning if they fight. Each side's expected value from war is roughly its chance of winning, minus what fighting costs it in blood, money and time. Is there a settlement both prefer to that gamble? There must be: any division giving each a little more than its war-expectation is preferred by both, and the room for such divisions is exactly as wide as the combined cost of fighting. The more destructive the war, the more deals both sides prefer to it.

So the range always exists. Why do rational parties fail to find and hold a point inside it? Two serious answers.

First, nobody can see where the range is. Its location depends on each side's real strength and real willingness to bear pain, and those are private. Would you simply tell your adversary? You would tell them you are strong and resolved regardless, because that shifts the settlement your way. But notice the trap: since exaggerating pays, your claim is discounted even when true. Both sides bargain with a distorted picture, and each may sincerely believe its demands sit inside the range when they do not. Hence credibility must be bought with actions a bluffer would not take — mobilizing at ruinous expense, making threats that would be humiliating to abandon. And hence fighting is itself informative: a war is, among other things, a very expensive way of learning what the other side is made of, which is roughly why wars end when they do.

Second, and harder: even a deal both sides want today may be unenforceable. Suppose power is shifting — one side will be stronger in five years. Any bargain struck now reflects today's balance, and the rising side will have reason to reopen it later. Can it promise not to? It can say so. Can it bind itself? Between states there is no authority above them to enforce it, so the declining side may prefer to fight now, while it still can. The same structure explains why civil wars are so much harder to end: a peace deal usually requires the rebels to disarm, and a disarmed party cannot enforce the promises made to it, so it may reasonably keep fighting rather than accept a bargain it cannot hold anyone to.

Three caveats, because a clean model invites over-reading. It treats each side as a single actor with coherent preferences, hiding much — leaders have domestic audiences, careers, and coalitions whose interests are not the state's. Second, "rational" here means only that choices follow consistently from beliefs; the beliefs may be badly mistaken, and misperception has an enormous historical record. Third, this is a framework, not a settled empirical finding, and how much of real war it explains is argued over.

The payoff survives all three. Once the peace-preferred range is visible, the question stops being "why did they want war" and becomes "what stopped them keeping a deal" — which points at fixable things: verification, third-party guarantees, transparency about capabilities, and mechanisms that make promises costly to break.

• Whether third-party peacekeepers and power-sharing guarantees really change civil war outcomes.
• Whether indivisible stakes — a holy site, a throne — are genuinely indivisible or made so by politics.
• How nuclear weapons alter the picture by making war's cost enormous and the balance hard to misjudge.

Every term the collection defines is gathered in the glossary.