Civics · Law · Geopolitics: The Mechanics and Dilemmas of Democracy

The core of an election is a translation problem: how do millions of votes convert into legislative seats? Change the rule and the result can differ wildly — the same public opinion can assemble entirely different governments. Two great families: plurality/majority systems (each district takes only the top finisher; winner-takes-all) and proportional representation (seats allocated in proportion to votes). The former wants "a clear winner and an accountable government"; the latter wants "a parliament that mirrors society to scale." There is no neutral translation — the rule itself shapes the outcome.

French political scientist Maurice Duverger proposed an empirical regularity in 1951: single-member plurality tends to produce a two-party landscape — voters, fearing a "wasted vote," gravitate to the two strongest parties. "Duverger's law" is no iron rule (later research found many exceptions), but it captures how rules in turn shape voter behavior.

Plurality's defenders: elections should produce a government that can be held responsible and get things done; voters should clearly know "who to blame if this term fails, and who to swap in next time." Under PR, responsibility is diluted across coalitions, so no one is accountable. PR's defenders reply: a party winning 40% of votes but 60% of seats, while another group is 20% of the population yet holds almost no seats, is itself unjust; parliament should be a mirror of society, not a stage for winners. The sacrifices are exact opposites — plurality trades representativeness for decisiveness; PR trades decisiveness for representativeness.

"The party with the most votes should naturally govern" — in many systems this is not so. In parliamentary systems, forming a government depends on whether you can assemble a majority coalition; the largest party, if no one will ally with it, can still end up in opposition. Under plurality, a party can even trail in total votes yet win more seats. Seats are not votes — this is a consequence of institutional design, not fraud.

Direct democracy (everyone votes on everything) is prohibitively costly and hard to make expert at large scale, so we get representation: a few are entrusted to decide on behalf of the many. But this immediately raises a fundamental question — once elected, should a representative strictly execute voters' wishes, or use their own judgment for the common good? This is the "delegate vs. trustee" dispute, a tension built into representation, not a bug that can be permanently removed.

• Trustee model: British statesman Edmund Burke's famous 1774 speech to the electors of Bristol — "Your representative owes you, not his industry only, but his judgement; and he betrays you instead of serving you if he sacrifices it to your opinion." He argued a legislator should answer to the whole nation, not to the opinion of one place at one time.
• Delegate model: treats the representative as the voters' mouthpiece, strictly executing the district's demands; referendums, binding party whips, and recall mechanisms all pull representatives toward being "executors."
• Party mediation: in reality legislators must both respond to voters and obey party discipline — a three-way tug-of-war among "voter–party–legislator," not the textbook binary.

The trustee model's strongest logic: voters often lack the time and information to judge complex issues (monetary policy, foreign treaties) and need representatives to apply expert judgment and answer for the long term. The delegate model's strongest logic: once a representative departs from voters' wishes, "representation" can degenerate into elites doing as they please, breaking the chain of authorization. Their costs run opposite — the trustee model may slide into elite paternalism detached from opinion; the delegate model may enslave politics to short-term mood, sacrificing the long term and minorities.

"If a representative doesn't vote my way, they've betrayed the voters" — this holds only under the delegate model. Representation was designed from the start to include room for "the representative's independent judgment"; exactly how much to follow opinion versus exercise judgment is a question of degree and a value choice, with no single correct answer.

Democracy's basic rule is "the minority submits to the majority," but the rule hides a danger: if the majority can decide everything, the minority's basic rights have no protection at all. Majorities can err too, and can oppress minorities for their own gain — this is the "tyranny of the majority." In the 19th century both Tocqueville and J.S. Mill warned that despotism need not come from a single tyrant; it can come from opinion and legislation swept along by a majority. So mature democracies commonly add a "fuse": some things cannot be done even if the majority agrees.

• Constitutional bill of rights: writing basic rights — speech, belief, the person — into the constitution, above majority vote, so even a majority cannot strip them away by legislation. This is a form of democratic self-limitation.
• Constitutional review: an independent court reviewing whether a law violates rights, in effect fitting a "circuit breaker" onto majority rule. The cost: a few judges can overturn the decision of an elected majority body, raising the retort "who guards the guardians?"
• Pitting faction against faction: American founder Madison argued in Federalist No. 10 for a large, diverse republic letting factions check one another, so that no single majority can dominate for long — offsetting majority tyranny.

A deep dilemma: does putting limits on the majority itself become anti-democratic? One camp holds that majority rule without rights protections self-destructs (a majority can vote to cancel the next election), so limiting the majority is precisely safeguarding democracy's foundation. The other camp worries that too many "counter-majoritarian" mechanisms (powerful courts, hard-to-amend constitutions) leave an elected majority unable to move, quietly transferring governing power to unelected judges and elites. The former sacrifices the purity of majority will; the latter sacrifices the safety of minorities and the long term — this is constitutional democracy's eternal tuning.

"Democracy is just the minority submitting to the majority" — this is only half right. Majority rule without a floor of rights is "illiberal democracy," which can legally oppress minorities; where the red line is drawn, and who guards it, is where countries truly differ.

Democracy is not a system that, once installed, runs forever; it needs maintenance and can degrade. Notably, today's democratic decline is often no longer a coup with tanks rolling into the square, but slow erosion from within by legally elected leaders — gradually weakening courts, co-opting media, rewriting electoral rules, each step seemingly within the legal framework, yet together hollowing out the institution. Scholars Levitsky and Ziblatt call this democracy's "gradual death": no clear moment of crossing the line, and by the time it's noticed, it is far gone.

• Erosion of informal norms: many key constraints live not in written law but in "conventions politicians tacitly observe" — restraint, not treating opponents as enemies, not using every legal means to annihilate them. Once this "software" fails, the "hardware" (constitutional text) cannot hold the system together.
• Differences in resilience: some democracies, hit by populist shocks, still steady themselves through redundancy — independent judiciary, professional civil service, decentralization, a strong civil society; others tilt fast because checks are thin. The difference often lies not in text but in institutional "depth."
• The corrosion of polarization: when society splits to where "losing to the other side equals catastrophe," each side grows more willing to break rules for victory — polarization is both cause and consequence of decline, a vicious cycle.

Defending democracy is itself fraught with tension. "Militant democracy" holds: anti-democratic forces (those inciting hatred or plotting subversion) should be restricted by law, lest democracy's tolerance pave the way for its gravediggers. Liberals are wary: the name "protecting democracy" is easily abused by incumbents to suppress legitimate dissent, so the restriction itself can become a tool of erosion. The former sacrifices some tolerance to hold the line; the latter bears the risk of being exploited in order to preserve tolerance.

"Once established, democracy won't backslide" — history refutes this again and again; institutions, like bridges, need ongoing upkeep. Another misconception is "as long as elections run on time, all is well": if the playing field has been quietly rigged, regular voting can instead provide legal cover for erosion. A democracy's health is read across the whole "authorize–constrain–replace" chain, not a single election day.