If so many revolutions devour their own, what about the ones that did not? The honest answer begins with a sobering fact: violent revolutions that produced stable, free governments are few. The American Revolution is the classic case, and even it excluded enslaved people, women and Native nations from the freedom it proclaimed. Many of the clearer successes since 1945 were negotiated or largely nonviolent. Looking across them, several common ingredients stand out.

Institutions to inherit

Revolutionaries who already know how to govern themselves have a head start. The American colonists had run elected assemblies, town meetings and courts for generations. When royal authority vanished, they inherited working institutions instead of inventing them, and the habits of debate, voting and accepting results were already familiar. Where a revolution has to build everything at once, from courts to tax offices, the temptation to rely on force is far greater. Political experience is a kind of capital, and societies with more of it tend to spend it well.

Limited aims

Revolutions that lasted usually had limited political aims rather than remaking people. Their leaders wanted to change who governed and how, to secure rights and to end particular abuses. They did not promise to abolish selfishness, create a new kind of human being or erase every inequality within a few years. Limited aims leave room for disagreement after victory. The American founders quarreled fiercely about tariffs, banks and the powers of the national government, but those quarrels ran through elections, newspapers and courts. A revolution that promises perfection will treat every failure as sabotage and every critic as an enemy.

Accepting loss

The surest sign that a new order has taken root is that its founders can lose power and hand it over. The political scientist Samuel Huntington proposed a two-turnover test: a democracy can be considered consolidated after power changing hands peacefully twice following its founding election. The party that won the first election must lose and leave office, and then its successors must do the same. The test is demanding, and many new democracies never pass it. But it captures something essential: freedom lasts only when losers believe they can win next time, and winners believe that losing is not ruin.

Pacts and bargains

In a landmark 1986 study, Guillermo O'Donnell and Philippe Schmitter examined transitions from authoritarian rule in southern Europe and Latin America. They noticed that regimes are rarely united: they contain hard-liners, who want to keep power at any cost, and soft-liners, who see that some opening is needed. Oppositions likewise contain moderates and radicals. Transitions were most likely to succeed when regime soft-liners deal with opposition moderates, often in explicit pacts that guarantee each side's core interests, such as protection for property, the armed forces or the losers' personal safety.

Spain is the best known example. After the dictator Francisco Franco died in 1975, his appointed successors steered a careful course. What was unusual about the 1976 Law for Political Reform was that Franco's own parliament voted to pass it, effectively voting itself out of existence. Free elections followed in 1977, and a democratic constitution in 1978. South Africa followed a similar logic in the early 1990s, when negotiators accepted temporary power sharing and job security for officials of the old state, known as sunset clauses, in exchange for majority rule. The scholar George Lawson called such cases negotiated revolutions: deep change reached through bargaining, not victory.

Pacts have critics. They can shield old elites from justice: Spain's 1977 amnesty law protected officials of the dictatorship from prosecution, and some victims' families still feel the past was buried rather than faced. Bargains can also lock in economic privileges that later governments struggle to touch. Defenders reply that a settlement which leaves some wrongs unpunished may still save many lives, and that justice can be pursued later through courts, archives and memory.

Worked example: South Africa, 1990 to 1996

In February 1990 President F. W. de Klerk lifted the ban on the African National Congress and other movements, and Nelson Mandela walked out of prison after twenty-seven years. Talks began, but violence surged: political killings, much of it between rival Black organizations and fueled by elements of the security forces, cost thousands of lives over the next four years. Negotiations broke down in 1992 after a massacre at Boipatong, then resumed. When the Communist leader Chris Hani was assassinated in April 1993, Mandela went on national television to appeal for calm, speaking with an authority the sitting government lacked.

The bargain held. The first fully democratic election took place in April 1994, a Government of National Unity included the old ruling party, and a permanent constitution followed in 1996. Then came a step that many revolutionary leaders never take: Mandela served one term and stepped down in 1999. Institutions to inherit, limited aims, a pact between moderates and a leader willing to give up power all appear. So do the costs: economic inequality largely survived the transition, which is why South Africans still argue about what the bargain achieved.