What Democracy Is — and What It Can and Cannot Do

We have studied power (11.1), the state that concentrates it (11.1.2), and the movements that contest it (11.2). Now we turn to the arrangement most of the modern world claims to live under, and the word almost everyone praises and almost no one defines: democracy.

Here is a fact that should unsettle you before we start. Nearly every government on earth calls itself a democracy — including many that hold no free election, jail their critics, and would be called tyrannies by any ordinary person. The word is so prized that even its enemies fly its flag. A word that everyone claims and no one can be refused is a word in serious danger of meaning nothing — and untangling what it actually means, what it can deliver, and what it cannot, is one of the most clarifying things this whole Part can do for you. Because you have opinions about democracy. Almost everyone does. And most of those opinions rest on a definition the holder has never examined.

The village that voted to take the well.

Imagine a village of a hundred families. Ninety of them share one ancestry, language, and faith; ten are different — a minority who have lived there for generations, farming a stretch of land with the village's best well.

The village is a perfect democracy in one sense: every decision is put to a vote, one family one vote, and the majority rules. No fraud, no coercion, no rigging. The purest majority rule imaginable.

One day the ninety vote to take the ten's land and their well, and divide it among themselves. The vote is ninety to ten. It is free, fair, and overwhelming. By the rule "the majority decides," the ten must hand over their land.

Now ask yourself a question that will not let you go: was that democratic? In one sense, unarguably yes — it was a free and fair majority vote, the very thing we praise. In another sense it is a horror, a tyranny of the ninety over the ten, exactly the kind of thing we thought democracy was against. Both answers are correct, and their collision is the deepest problem in the theory of democracy. The vote was democratic as a procedure and monstrous as an outcome — and if your definition of democracy is only "the majority decides," you have no way to say what is wrong with taking the well. You have to reach for something the vote alone cannot give you. Holding that thought, we can begin.

Two definitions, pulling in opposite directions

Almost every confused argument about democracy comes from two people using the same word for two different things. There are two great families of definition, and they answer different questions.

The minimalist (procedural) definition — democracy as a method. On this view, associated with Joseph Schumpeter, democracy is not about outcomes at all; it is simply a method for choosing rulers: a system in which those who govern are selected through free, fair, competitive elections, and can be removed by the same means. That is the whole of it. Democracy, minimally, is "a political method... an institutional arrangement for arriving at political decisions in which individuals acquire the power to decide by means of a competitive struggle for the people's vote." The people do not rule; they choose and dismiss those who rule. Its virtue is not that it produces wise or just policy but something humbler and huge: it lets a society change its rulers without bloodshed.

The substantive (participatory) definition — democracy as a condition. On this richer view, democracy is not just a method but a whole way of life and set of outcomes: real participation by ordinary people in the decisions that affect them, genuine political equality, an informed and engaged citizenry, and often social and economic conditions (education, security, some equality of resources) without which the vote is hollow. On this view a country with free elections but a passive, misinformed, powerless population, where a tiny elite really decides everything (recall 11.1.3), is at best a thin and impoverished democracy — democratic in form and barely in substance.

Notice these are not just two definitions but two values, and each catches something the other neglects. The minimalist definition is modest and testable — you can check whether elections are free, fair, and competitive — and it guards the one achievement no one should sneer at: the peaceful removal of rulers. But it can bless a hollow, elite-run system as fully "democratic" so long as the voting ritual is clean. The substantive definition is demanding and inspiring — it insists democracy means real popular power, not a ritual — but it is so demanding that no actual country fully meets it, and it can be stretched to justify almost anything ("real" democracy requires my preferred outcome, so overriding your vote is the truly democratic move). Keep both in view: the minimal one to protect the irreducible core, the substantive one to keep you from mistaking the ritual for the reality.

What democracy actually is, in practice: polyarchy

Because full substantive democracy exists nowhere and pure proceduralism feels too thin, the political scientist Robert Dahl (whom we met governing New Haven in 11.1.3) offered a realistic middle term for what the best real systems actually are. He refused to call them "democracies" — that word, he thought, names an ideal no country reaches — and coined polyarchy for the real thing: rule by the many, as opposed to rule by one or few.

A polyarchy, in Dahl's account, is a real political system that comes closest to the democratic ideal, marked by a bundle of institutions that must all be present:

elected officials who control government decisions;
free, fair, and frequent elections;
inclusive suffrage — nearly all adults can vote;
the right to run for office — nearly all can stand;
freedom of expression — to speak, including against the government;
alternative sources of information — a free press and independent sources, not a single official truth;
associational autonomy — the right to form independent parties, unions, and groups.

The point of the list is that democracy-in-practice is not just elections. Voting without free speech, a free press, the right to organise, and real alternatives on the ballot is a shell — because the vote only means something if the things that feed it (information, organisation, genuine choice) are free too. This is why regimes that hold elections but muzzle the press and ban real opposition are not "flawed democracies" but something else wearing the costume. The election is the visible tip; the freedoms that make it meaningful are the submerged mass, and Dahl's list is really a list of those.

The tension at the core: democracy versus liberty

Now back to the village and the well, because it exposes the deepest thing in this lesson. What was wrong with taking the well was not that the vote was unfair — it was perfectly fair. What was wrong is that a fair vote can produce a tyranny. And this reveals that the thing we call "liberal democracy" is actually two different ideas bolted together, and they are in permanent tension.

Democracy — in its core sense — means rule by the people, which in practice means majority rule: the larger number decides.

Liberalism — a separate idea — means limits on power, including the power of the majority: certain rights (of speech, of property, of the person, of minorities) that cannot be voted away no matter how large the majority.

These pull against each other. Pure majority rule says: whatever the majority decides, goes — including taking the well. Liberalism says: no, some things are off the ballot entirely — the majority may not vote to enslave the minority, silence dissent, or seize the well, however many they are. The "liberal" in "liberal democracy" is precisely the constraint on the "democracy." It is the part that answers the village: the ten's right to their land is not up for a vote, and ninety votes cannot make it so.

Neither half is safe alone — and this is not a flaw to be fixed but a balance to be held.

Democracy without liberalism is the tyranny of the majority — the village and the well, free and fair and monstrous. A majority, unconstrained, can vote to crush a minority, and does, throughout history. Elections alone do not protect anyone; they can be the instrument of oppression as easily as the guard against it. This is "illiberal democracy": real majorities, real elections, no limits, and the systematic crushing of whoever is outvoted.

Liberalism without democracy is rule by an unelected few who respect some rights but answer to no one — an enlightened despotism, or a technocracy, or a judiciary that overrides every popular decision. Rights are protected, but the people have no say; and an elite that grants rights can withdraw them, because nothing but its own goodwill compels it.

The genius — and the permanent instability — of liberal democracy is that it tries to hold both at once: the majority rules, and there are limits the majority cannot cross. This is not a problem to be solved but a tension to be managed forever, and every real system sits somewhere on the seesaw, forever contested. Push all the way to "the majority decides everything" and you get the well taken; push all the way to "rights override every vote" and you get rule by judges and experts over a people with no say. Most real political fights you will ever see are, underneath, fights about exactly where on that seesaw to sit — how much the majority may decide, and how much is placed beyond its reach. Seeing that is worth more than any slogan about "more democracy" or "protecting rights," because it shows you that both sides are usually defending one real value against another.

What democracy cannot do: the limits even friends must admit

To defend democracy honestly you must be honest about what it cannot deliver — because overselling it sets people up for the disillusion that its enemies feed on.

It does not guarantee good or wise decisions. Majorities can be wrong, ill-informed, short-sighted, or cruel. Democracy is a method for choosing and removing rulers peacefully, not a machine for producing correct policy. Its friends who promise it will always decide wisely are lying, and the lie backfires when a majority decides badly, as majorities sometimes do.

It cannot perfectly translate "what the people want" into policy — because there may be no coherent "what the people want" at all. This is a genuinely deep result. When more than two options are on the table, there can be no fair voting rule that reliably turns individuals' consistent preferences into a consistent group choice: majorities can cycle (the group prefers A to B, B to C, and C to A), and the "winner" can flip depending on the order things are voted on or which method is used. This is not a fixable defect in one voting system; it is a limit on the very idea of a single "popular will." "The will of the people" is often not a thing that exists to be found, but an artefact of the rules used to aggregate — which means whoever controls the rules (the agenda, the method) shapes the outcome (the second face of power, 11.1.3, now at the heart of democracy itself).

It does not by itself produce economic equality, competence, or even stability. Democracies vary enormously on all of these; the vote guarantees none of them. What it reliably provides is narrower and more precious than any of them, which is the next point.

What democracy can do: the modest, decisive achievement

Set against those limits, what democracy does deliver is easy to undervalue precisely because we are used to it — and it is worth stating plainly, because it is the honest case for democracy once the overselling is stripped away.

It lets a society remove its rulers without killing them. This sounds small and is colossal. For almost all of human history, the only way to change a bad ruler was death — theirs or yours: assassination, revolt, civil war, invasion. Democracy's core achievement is a procedure for peaceful succession: a way to say "you're fired" to the most powerful people in the land, and have them actually go, without a shot. A system in which those in power can be voted out, and leave, is doing something most political systems in history could not do at all — and it does it precisely by making power temporary and revocable, which is the single strongest check on its abuse, because a ruler who can be removed has reason to fear the ruled.

And downstream of that one achievement come others, well supported by evidence: governments that can be voted out are, on average, more responsive to ordinary people's basic needs, less prone to the very worst catastrophes (large-scale famines, for instance, are strikingly rare in functioning democracies with a free press, because rulers who can be removed and criticised cannot ignore mass starvation the way unaccountable ones can), and more able to correct their own mistakes — because the mechanism that removes rulers also, over time, removes policies that fail. Democracy is not a promise of good government. It is a mechanism for getting rid of bad government cheaply — and that, not any guarantee of wisdom, is the honest heart of the case for it.

Two countries, one famine.

Consider two countries hit by the same harvest failure. In the first, the rulers cannot be removed and the press cannot report freely. Officials have every incentive to hide the shortage, deny it, and protect themselves; the news does not travel; relief does not come; people starve in the dark while the capital insists all is well. In the second, rulers face an election and a free press. The shortage is reported at once, opposition politicians seize on it, the government knows that letting people starve means being thrown out — so relief moves, because not moving it is politically fatal.

The point is not that the second country's leaders are kinder. It is that democracy changes their incentives: a ruler who can be removed, in a country where the failure can be reported, cannot afford the indifference an unremovable ruler can. This is the case for democracy at its strongest and most honest — not that it makes leaders good, but that it makes their survival depend on not being catastrophically bad, and gives the ruled a bloodless way to enforce that. It is a modest claim, and it has saved an immense number of lives.

Why this matters

Because "democracy" is going to be invoked at you, constantly, by people who want your assent — and most of them are relying on you never having distinguished the method from the condition, the majority from the limits, the ritual from the reality.

When someone says a country "is a democracy" because it holds elections, you now ask: and the free press? the real opposition? the rights the majority cannot vote away? — Dahl's submerged list, without which the election is a costume. When someone says a policy is legitimate because "the majority wants it," you now ask two things: is there really a coherent majority will here, or an artefact of the voting rule? and is this one of the things a majority may rightly decide, or one that should be off the ballot? — the well and the seesaw. And when someone attacks democracy for failing to produce wise or equal or prosperous outcomes, you can grant the point honestly — it guarantees none of those — while holding the one thing it does guarantee, which no rival system offers: the peaceful, bloodless removal of the powerful by the powerless.

That clarity is a kind of citizenship. Most defences of democracy oversell it, and most attacks on it exploit the overselling. The sturdy position — the one that survives contact with democracy's real failures — is the honest one: it does far less than its friends promise and far more than its critics admit, and the far-more is the one thing that matters most. This is the same evenhandedness this course demands everywhere (Part 7): claim for democracy exactly what the evidence supports, no more, and defend that — because a case built only on the truth cannot be knocked down by pointing at the truth.

The next lesson (11.3.2) takes up the "discontents" in this topic's title directly: the strains showing up in democracies across the world at once — declining trust, polarisation, the sense that the vote has come unhooked from real power — and asks, with the same honesty, which of these fears the evidence supports and which it does not.

What to carry out of this lesson:

"Democracy" hides two definitions. The minimalist/procedural (Schumpeter): a method for choosing and peacefully removing rulers by free, fair, competitive election. The substantive/participatory: real popular power, political equality, an informed engaged citizenry. The first is modest and testable; the second inspiring but met fully nowhere.
Real systems are polyarchies (Dahl), defined by a bundle — elections plus free expression, a free press, the right to organise, real alternatives. The freedoms that feed the vote are what make it mean anything; elections without them are a costume.
Liberal democracy bolts together two ideas in permanent tension: democracy (majority rule) and liberalism (limits the majority cannot cross). Democracy without liberalism is the tyranny of the majority (the village and the well); liberalism without democracy is unaccountable rule by the few. The balance must be held, not solved.
Democracy's honest limits: it does not guarantee wise or just decisions; it cannot perfectly translate "the popular will" into policy (there may be no coherent will — majorities can cycle, and the rules of aggregation shape the outcome); and it does not by itself deliver equality, competence, or stability.
Democracy's decisive achievement is modest and immense: it lets a society remove its rulers without bloodshed, making power temporary and revocable — and downstream of that come responsiveness, the avoidance of the worst catastrophes, and self-correction. Not a promise of good government; a cheap mechanism for ending bad government.

Minimalist (procedural) democracy — Schumpeter's view of democracy as a method for selecting and removing rulers through free, fair, competitive elections, judged by the process, not the outcomes.

Substantive (participatory) democracy — the richer conception of democracy as real popular participation, political equality, and an informed citizenry, often requiring social and economic conditions; demanding, and fully realised nowhere.

Polyarchy — Dahl's term for real systems that best approximate democracy, defined by a bundle of institutions: elections plus inclusive suffrage, free expression, a free press, and associational autonomy.

Liberalism (in "liberal democracy") — the principle of limits on power, including certain rights that cannot be voted away by any majority, however large; the constraint on majority rule.

Tyranny of the majority — the danger that a free and fair majority can use its power to oppress a minority; why democracy without liberal limits can be an instrument of oppression.

Aggregation problem / cycling — the result that, with more than two options, no fair voting rule reliably turns consistent individual preferences into a consistent group choice, so "the popular will" is often shaped by the voting rules rather than simply revealed by them.

Peaceful succession — democracy's core achievement: a procedure for changing rulers without violence, making power temporary and revocable.

Try it yourself:

One — Which democracy do they mean? Next time you hear a country praised or damned as "a democracy" or "not really a democracy," work out whether the speaker is using the minimalist definition (are the elections free and competitive?) or the substantive one (is there real popular power, equality, an informed public?). Many disputes dissolve once you see the two sides are scoring the same country against different definitions.

Two — Put it on the seesaw. Take a live political controversy — a court striking down a popular law, a referendum on a minority's rights, a leader claiming a mandate to override the courts. Locate it on the democracy-versus-liberalism seesaw: is this a case of majority rule pressing against limits on the majority? Name what each side is really defending. You will usually find both are defending a real value against another real value — which is a truer picture than "one side is against democracy."

Three — Test the will. Think of a group decision among three or more options you have witnessed (a committee, a friend group choosing where to eat). Notice how much the outcome depended on the procedure — who set the shortlist, what got voted on first, whether it was one round or a run-off. That dependence is the aggregation problem in miniature: the "group's choice" was partly made by the rules, not just the preferences — and whoever controlled the rules held a quiet power over the result.