Democracies are not saved by citizens who never disagree, never lose, never feel threatened, and never doubt. They are saved by citizens who can experience all of those things without withdrawing from the common project.

Democracies rarely die because everyone suddenly believes in authoritarianism. They more often weaken because too many people who know better decide, one by one, that saying so is not worth the cost.

A citizen sees a lie hardening into a party line and stays quiet. A public official knows an election was legitimate but softens the sentence. A donor dislikes the demagogue but likes the tax cut. A neighbor hears a friend speak with contempt and changes the subject. A judge follows the law, but not too loudly. A journalist knows the public record is being poisoned but still treats the poison as merely one side of a debate.

None of these choices feels decisive. Each has a reasonable explanation. People have families, jobs, congregations, clients, donors, reputations, futures. They do not experience themselves as cowards. They experience themselves as prudent. But democracy is built out of these small acts of prudence. Enough of them, repeated long enough, can make a country look as though it has changed its mind when what has really changed is the cost of speaking honestly in public.

The central question for American democracy, then, is not only whether the institutions will hold. It is whether enough citizens believe they will not be alone if they tell the truth.

Call this the courage threshold.

The phrase borrows from a body of work on preference falsification, social cascades, and collective action associated with thinkers such as Timur Kuran and Cass Sunstein. Kuran's central insight in Private Truths, Public Lies is simple and devastating: people often misrepresent what they privately believe because public honesty carries social costs. When enough people do this, society loses track of itself. The public record no longer records public belief. It records public fear.

Sunstein's work on social norms and cascades adds the mechanism. People do not need identical levels of courage. They have different thresholds. Some will speak when no one else does. Others will speak when ten people have. Others need a crowd. In moments of change, what matters is not merely what people believe, but what they believe other people believe — and whether they can see enough others moving first.

This is why public life can seem frozen until it suddenly is not. A regime can look stable the day before it collapses. A social movement can look marginal the week before it becomes obvious. A taboo can look permanent until enough people break it at once. The hidden variable is not belief. It is perceived company.

That is the part of democratic life we under-measure. We count votes, poll opinions, track trust in institutions, chart polarization, and monitor misinformation. All of that matters. But democracy also depends on whether citizens believe they can bear the ordinary costs of citizenship — disagreement, sacrifice, loss, embarrassment, social friction — without being abandoned.

Democracy is not only a system for aggregating preferences. It is a system for making sacrifice legitimate.

That is one of Danielle Allen's most useful contributions to democratic theory. In Our Declaration and Justice by Means of Democracy, Allen returns democracy to an older moral terrain. Self-government is not merely the exercise of rights. It is the practice of sharing burdens. Every policy allocates pain. Every law asks someone to give something up. Every democratic settlement tells some citizens: you lost this round, but the process still belongs to you.

That bargain is hard enough under normal conditions. It becomes nearly impossible when citizens experience loss not as a fair defeat, but as humiliation by an enemy group.

Here Lilliana Mason is indispensable. In Uncivil Agreement, Mason shows that American polarization is not simply a matter of citizens holding more extreme policy positions. The deeper problem is social sorting. Party identity has absorbed race, religion, geography, education, class, culture, and lifestyle until politics has become a proxy for who people are. When that happens, losing an election no longer feels like losing an argument. It feels like status subordination.

Democracy can survive disagreement. It struggles when disagreement becomes identity defense.

This is where institutional accounts of democracy often become too tidy. We like to say that democracy depends on rules, checks and balances, independent courts, fair elections, parties, procedures, norms. It does. But the machinery assumes a citizenry willing to keep using it even when it hurts.

James Madison understood that faction was permanent. The Constitution did not eliminate conflict; it organized it. The point was not to make Americans agree, but to make disagreement governable. Fareed Zakaria's long-running concern with liberal democracy, institutions, and culture belongs in this tradition, as does Daron Acemoglu and James Robinson's The Narrow Corridor. Liberty survives only in the corridor where the state is strong enough to govern and society is strong enough to restrain the state.

But corridors are not self-maintaining. Formal design creates the arena. Norms determine whether people keep playing when they lose.

The United States has never been as good at this as its self-image suggests. One reason Jill Lepore is so useful in this conversation is that she resists the temptation to tell American history as either inevitable progress or inevitable betrayal. In These Truths, This America, and her writing for The New Yorker, Lepore returns again and again to the asterisk in the American promise. Equality declared, exclusion practiced. Consent celebrated, coercion hidden. Inquiry promised, trust eroded.

American democracy has nearly died before. It has also been born later than Americans like to admit. If democracy means meaningful political equality across race and sex, then the United States is not almost 250 years old. It is closer to sixty. The Civil Rights Act and Voting Rights Act did not perfect American democracy, but they made its founding claims more nearly true than they had ever been.

That should make us both more hopeful and less complacent. More hopeful, because democratic breakthroughs are possible even after long injustice. Less complacent, because the full democratic bargain is younger, thinner, and more reversible than the national myth implies.

A democracy this young cannot afford to misunderstand silence.

Silence is often treated as consent. In politics, it is frequently nothing of the kind. Silence can be exhaustion. It can be fear. It can be the rational response to an environment in which every sentence becomes evidence in a social prosecution. It can be what happens when citizens no longer know which institutions will protect them, which neighbors will stand with them, or which facts will still be admissible tomorrow.

This matters because democratic resilience depends on public feedback. Institutions need to know when they are failing. Parties need to know when they have crossed lines. Movements need to know whether their claims resonate beyond their loudest members. Citizens need to know whether their private doubts are shared. If preference falsification becomes widespread, every part of the system starts operating on bad information.

Markets fail when prices stop carrying information. Democracies fail when public speech stops carrying belief.

This is the bridge between the democracy crisis and the epistemic crisis. We often talk about misinformation as though the problem were simply falsehoods moving through a population. But a democracy can be damaged just as badly by true beliefs that cannot move. When citizens privately doubt a lie but publicly repeat it, the lie gains power. When citizens privately reject extremism but publicly accommodate it, extremism gains legitimacy. When citizens privately want a different politics but publicly perform the politics their group expects, the system mistakes performance for preference.

That is why debate boundaries matter so much in a post-consensus society.

A healthy democracy needs disagreement sharp enough to expose error but bounded enough to preserve shared membership. The old civic fantasy was that public debate worked like a marketplace: arguments entered, citizens evaluated them, better claims won. That model depends on conditions we no longer reliably possess: trusted sources, low identity stakes, shared standards of evidence, and enough cross-cutting relationships to keep opponents human.

Once those conditions erode, argument itself changes character. Evidence becomes ammunition. Correction becomes insult. Compromise becomes betrayal. Public reasoning becomes less a search for truth than a performance of loyalty.

The answer cannot be to give up on argument. Democracies need argument. But we should be more honest about what argument can and cannot do when citizens are sorting themselves into identity camps. Before evidence can persuade, people need enough trust to hear it and enough security to survive changing their minds.

That means democratic repair is not only institutional reform. It is social reconstruction.

Some of that work is procedural: open primaries, ranked-choice voting, independent redistricting, voting rights protections, local journalism, civic education, and stronger guardrails against election subversion. Procedures matter because incentives matter. Bad systems reward bad behavior and then blame citizens for responding predictably.

But some of the work is pre-political, or at least pre-argumentative. Citizens need places where they can encounter one another without every encounter being a referendum on national identity. They need institutions that lower the cost of honesty: churches, unions, veterans' groups, local newspapers, schools, neighborhood associations, civic assemblies, and parties that act like parties rather than content brands. They need leaders who understand that the point of rhetoric is not merely to mobilize the already convinced, but to make courage less lonely.

The great democratic speeches do this. They do not merely inform. They alter the perceived company of the listener. They make private conviction feel public enough to act on.

Frederick Douglass did this. Abraham Lincoln did this. Martin Luther King Jr. did this. Barbara Jordan did this. They named the country's failures without exiling the listener from the country's future. They made courage contagious.

That may sound sentimental. It is not. It is a theory of social change.

Movements win when they change what people think other people are willing to say and do. Authoritarians win the same way. The difference is the content of the courage being summoned. A democratic movement asks people to bear the cost of truth, pluralism, and mutual obligation. An authoritarian movement asks people to bear the cost of domination and call it strength.

The American danger is not that no one sees what is happening. Many people see it. The danger is that too many people believe too few others see it, or that too few others will act if they do.

That is the courage threshold. Lower it, and democracy becomes resilient. Raise it, and democracy becomes brittle.

The practical question is therefore not only "What do citizens believe?" It is "What would they be willing to say if they knew they were not alone?" Not only "Do they support democracy?" but "What cost will they bear for it?" Not only "Do institutions still exist?" but "Do citizens still experience those institutions as worth losing through, sacrificing for, and telling the truth inside?"

A democracy is not saved by citizens who never disagree, never lose, never feel threatened, and never doubt. It is saved by citizens who can experience all of those things without withdrawing from the common project.

The next phase of democratic repair will require many reforms. But beneath them is a simpler and harder task: make honest citizenship less lonely.

Because democracy does not only depend on what people privately believe. It depends on the moment when enough of them discover that their private belief has public company.

Reading / sourcing

On preference falsification and social cascades

On democracy as shared sacrifice

On polarization as identity sorting

On institutions and the narrow corridor

On American democratic history

By Ted Delicath. Published on adambeday.org, 2026.