Indexing the archive…
Your Universe of Digital Possibilities
Design a fair way for nine people to choose among three candidates. Weight the ballots how you like — plurality, Borda, a runoff, head-to-head — and the machine will find, in your own electorate, a case that breaks independence: two elections where nobody changed their mind about the top two, only an irrelevant third candidate moved, and the winner flipped. Three consistent voters can already make a majority that circles (Condorcet, 1785); Arrow proved in 1951 that no ranked rule escapes it for three or more options. The one door is exact — two candidates, and majority rule is perfect and unique (May, 1952). This instrument ends at the wall, on purpose.
Perfectly consistent individuals can build a society that is not: three voters ranking α>β>γ, β>γ>α, γ>α>β give a majority for α over β, for β over γ, and for γ over α — a preference that eats its own tail. There is no Condorcet winner; the circle was there before any rule.
For three or more options, no rule that turns individual rankings into a social ranking can honour all three of unanimity (if all prefer x to y, so does society), independence of irrelevant alternatives (the x-vs-y verdict depends only on x-vs-y opinions), and non-dictatorship. Every rule surrenders one — or crowns a dictator. Not hard: impossible.
Shrink the choice to two candidates and the wall opens exactly: simple majority rule is the one and only method that is anonymous (voters interchangeable), neutral (candidates interchangeable) and monotone (more support never hurts). With two options fairness is not just possible but forced — the door Arrow’s wall leaves standing.
The cycle came first: Condorcet saw in 1785 that consistent individuals can vote a society into a circle, α over β over γ over α, with no one at the bottom. For a century that looked like a curiosity. Arrow made it a wall in 1951 — no ranked rule on three or more options can honour unanimity, independence and non-dictatorship together; surrender one or crown a dictator. This is the cycle’s clearest case of hard versus impossible: it is a theorem, closed and dated, not a cynicism. Pairwise dynamics — who-beats-whom as a tournament — live elsewhere on the rack in The Tournament; Noa owns the impossibility itself. The one door is exact: with two candidates, May proved majority rule is not just fair but unique. As with every wall in this edition, the impossibility is the finding.