Chapter 1 called the modern ballot a one-bit signal. Here is what a wider channel looks like — and why the width is not a luxury but the difference between a system that hears its citizens and one that merely counts them.
Five voters face a field of candidates: A and the others. Two of the five like A. Three strongly oppose A — but they're split on which of the others they'd prefer instead. Under binary voting, A wins — two marks against a scattered one, one, one — and the opposition's intensity counts for nothing because a mark has no magnitude. The outcome is exactly backwards from what the group actually wants: a minority's mild preference defeats a majority's strong objection, because the channel couldn't carry the word strong.
Now give each voter a budget of points, spendable for or against anyone on the ballot. The two supporters put +1, +1 on A; they like A, but not enough to spend more. The three opponents put −2, −2, −2 — they don't need to agree on an alternative to register that this matters to them. Net: −4. A loses. Nothing exotic happened; the system simply translated political will correctly, because the signal carried three dimensions instead of one: engagement (did you spend attention on this at all), direction (for or against), and magnitude (how much of your scarce budget it was worth to you).
Each citizen receives a modest point budget per cycle — and so does each delegate: the same three-dimensional channel runs the public's agenda and verdicts and the delegates' passage votes alike. The budget forces prioritization — you cannot scream maximum intensity at everything — which means every allocation is information. Indifference finally shows up in the data: proposals that fail to draw engagement simply don't pass, no matter how their small audience leans, which is the filter that keeps a passionate clique from legislating for an indifferent public. And priority emerges by consent: the questions that gather the most backing are, definitionally, the ones the public most wants worked on. Taiwan's digital democracy experiments, the vTaiwan consultations that map where citizens already agree instead of tallying sides, have been circling exactly this insight — surface what has broad support behind it, rather than whatever wedge issue splits 51/49. A bare-majority win on a low-engagement question is not a mandate; it's an artifact of the old channel.
Quadratic voting — the economists Glen Weyl and Eric Posner's scheme where each additional vote on a question costs the square of the count, so two votes cost four and three cost nine — deserves its nod here: it is the best-known academic answer to intensity capture, the problem of registering how much a voter cares and not merely which way they lean, and convex cost curves of that family can serve as an optional damper within point-voting. This design doesn't adopt it here — not because quadratic voting is wrong, but because the two tools lean different ways. Its efficient-market framing — the idea that a price should absorb new information the instant it arrives — prizes fast preference-response; this proposal bets instead on deliberated, sustained judgment, and captures intensity through the point budget and engagement floors without the market metaphor. Both sit in the toolbox of chapter 2: quadratic voting stays a fully valid instrument a jurisdiction could adopt, and the search — not this document — gets the last word on where each tool fits.
Voting here is not an event; it is a window. Anything entering a voting cycle must have been submitted the cycle before — a built-in deliberation lag that kills midnight riders — unrelated provisions slipped into a bill in its final hours — and surprise votes. Points are allocatable and freely movable throughout the window, locking only at the boundary, so early voters are never punished by information that emerges mid-window, and a vote cast five minutes before the boundary weighs the same as one cast five days before. Batching decisions to a shared cadence — the governance tickrate — is what keeps a fast digital system deliberate at human speed: accumulate during the window, resolve at the boundary, and never let "who can sustain attention longest" substitute for "what has genuine support."
Passage is the delegates' arena. Chapter 4 seats them for exactly this: they spend their point budgets on the questions they judge important — professional attention where public attention is thin, which is what keeps the boring-but-vital, pension solvency and grid maintenance, from starving in an attention-gated system. Some questions never belong to them — the meta layer and the standing verdicts are the citizens' directly — and contested passages can be kicked back to the broad ballot. The rules below bind in either arena.
Passing requires a floor and a majority. The majority is net: weight for minus weight against, among the weight actually placed. The floor is affirmative: the winning option must gather a required share of all eligible weight — backing measured against society, not against whoever showed up.
The distinction sounds technical and is anything but. Participation quorums count presence, so absence becomes the strongest vote in the system: opponents learn to stay home, because a boycott kills quorum more reliably than a no-vote — Italy's referendum system died of exactly this. An affirmative floor deletes the strategy. Absence cannot block, because the floor counts only backing, which no boycott reduces; and absence cannot consent, because the floor is measured against everyone, so the silent majority weighs the scale down rather than being presumed to agree. The opponent's best move is always the honest one — show up and spend against. Silence never counts as a vote, for either side.
The floor is also where this design meets the tyranny of the minority. An intensity-weighted channel exists to let a committed few be heard — and network studies of committed minorities find that a bloc on the order of one voter in ten, unshakable and coordinated, can flip what a whole population does. A point system without a floor would institutionalize that: eleven percent, spending hard, out-voting a soft and scattered everyone-else on question after question. The affirmative floor is the dial that bounds it: it sets, explicitly and in the meta-rules, how large a share of society must actively back a policy before it binds everyone, and the risk-proportional friction of chapter 2 turns that dial up with a policy's blast radius. An intense minority can still put a question on the agenda and win the argument; it cannot govern from the indifference of everyone else. And whatever does pass narrowly lives under the citizens' standing verdict (chapter 8) — the one layer no minority can capture.
None of this presumes a yes/no ballot. A question may carry live options — A or B or neither — and the rules generalize cleanly: the net is computed per option, and the floor belongs to whichever option would win. Binary is one shape of ballot here, not the system.
Bundles get one honest concession: items may pass together as a negotiated package when their budgets only reconcile jointly — but they decouple legally the moment they pass. Severability is the default; each atom thereafter lives or dies on its own record. Omnibus-forever — everything permanently fused into one bill nobody can vote on piecewise — is how attribution dies (chapter 14).