Why do regimes seem rock-solid until the day they collapse?

Because what people say in public can be very different from what they think in private, and nobody can see the gap. Economist and political scientist Timur Kuran gave this a name in a 1987 article: preference falsification, meaning publicly expressing a preference that differs from your private one, under perceived social pressure. When many people do it at once, a collective illusion can form: most of a group goes along with an idea they don't agree with, because they incorrectly believe that most others agree.

Kuran's central claim is about what happens next. People differ in how much social incentive it takes to make them speak out; he calls those switchover points political thresholds. Until a critical number of people reach theirs, changes in private opinion are invisible to outsiders, even to one another. When private and public opinion are far apart, a shock of the right kind can push a critical number of disgruntled people past their thresholds and set off a public-preference cascade. He compares silent discontent to underground stresses building for decades without shaking the ground, and the sudden change to an earthquake.

The case Kuran uses is 1989. The collapse of several communist regimes in Eastern Europe surprised scholars, statesmen, intelligence agencies, dissidents such as Havel and Solzhenitsyn, and even Gorbachev. On Kuran's account, few would step forward while the opposition was minuscule, so no one knew how ready the region was. When East Germany's police responded with restraint during Gorbachev's visit of 7 October 1989, demonstrations grew explosively, and the Berlin Wall fell on 9 November.

Kuran adds a warning: good hindsight does not imply good foresight. Explaining a revolution afterwards does not make the next one predictable.