Nobody sits down and designs a failing institution. Institutions fail in patterned, predictable ways — the same ways, across centuries and continents — and any serious governance design has to start by naming those patterns. This chapter is the list of failure modes the rest of the proposal exists to avoid. Everything that follows is an answer to something on this page.
An institution is created to solve a problem. It hires, builds process, accumulates budget — and somewhere along the way its goal quietly changes from solving the problem to continuing to exist. From there it faces the same two deaths an organism does. The first is senescence: repair stops keeping up, errors accumulate, flexibility calcifies — Mancur Olson, the economist who traced how prosperous nations seize up, mapped this one as institutional sclerosis, the slow encrustation of interest groups and rules until the body can no longer respond to its environment. The second is cancer: a part of the body whose growth decouples from the body's function — the whitepaper's bureaucratic cancer, administration expanding past functional necessity, consuming resources while producing less, and evading every kill signal sent at it. Michael Levin, the biologist who studies how cells coordinate into bodies, gives the cellular version its precise name: the cancer cell has suffered a collapse of its cognitive light cone — cut off from the signals that bound it into the organism, it stops perceiving itself as part of anything larger and reverts to the only goal left to it, its own proliferation. The tell is in the cure: restore the signalling and such a cell can rejoin the whole. The institutional version runs the same course, and chapter 2 argues this is structure rather than rhyme. Both deaths are reliable because nothing in the institution's design ever scheduled its death or armed its immune system. Founding is ceremonial; sunset is unthinkable. So failure gets reframed as underfunding, headcount becomes a proxy for importance, and the org chart survives every mission it outlives. The whitepaper puts the sharpest possible point on it: a passive tool wants its users to see clearly; only a selfish agent generates opacity to prevent its own sunset. Institutional illegibility is not an accident of complexity — it is a survival strategy, selected for.
This is why the whole proposal is framed as institutional lifecycle management: how you spin an institution up, give it credibility, hold it to its purpose while it runs, and retire it once its work is done — with every one of those stages designed, not just the first. And it is why the design's counters are biology's counters, one to one: senescence is met with scheduled renewal — the reviews of chapter 8, repair on a cadence instead of decay by default; cancer is met with programmed death and immune surveillance — sunset clauses are the institution's apoptosis, and the capture bounties of chapter 14 — rewards paid for exposing an institution captured by the interests it governs — are its antibodies. An organism survives by running both defenses continuously. So does a government.
Modern suffrage asks each citizen one question every few years: which of these, once? That is a breathtakingly narrow channel for steering a civilization — direction with no magnitude, no intensity, no way to say "I don't care about this but I desperately care about that." As the whitepaper's specification puts it, we are steering with a signal that says left or right once every four years, with no indication of how much or why. Every pathology of modern politics that gets blamed on voters — polarization, single-issue capture, the tyranny of the marginal swing district — is partly an artifact of the channel being too narrow to carry what people actually think. Chapter 6 is about widening it.
Goodhart's law — when a measure becomes a target, it ceases to be a good measure — is not an occasional bug of governance; it is the default outcome of governing by metric. Test scores rise while education falls. GDP grows while the felt economy rots. Crime statistics improve while crime is reclassified. The deeper version, which the whitepaper returns to again and again, is about power: whoever controls the rubric controls the population. Any fixed, named metric is an invitation to the people it measures to capture it. Chapter 3 describes the one class of signal that resists this — and chapter 14 is honest about where the gaming relocates when you use it.
The principal–agent problem in one sentence: the people who govern (agents) do not carry the outcomes of the people they govern for (principals). A leader's payoff arrives at the next election; a policy's consequences arrive over decades. The whitepaper calls this the short light cone of political selection — leaders capture the upside of visible action now and diffuse the downside onto whoever holds office when the bill comes due. No amount of electing better people fixes a payoff structure that selects for short horizons. Chapter 7 rewires the payoff instead.
Peter Turchin's cliodynamics, the statistical study of why societies cycle into instability, adds the demographic driver: societies reliably overproduce elite aspirants relative to elite positions, and the surplus doesn't evaporate — it lodges in whatever institutions will hold it, competing ever harder over fixed seats, converting positive-sum institutions into zero-sum arenas. Elite overproduction is where sclerosis, capture, and bureaucratic growth get their energy. A governance design that ignores it will be colonized by it; this one meets it with fixed position counts, rotation, sunset of unused positions, and — most importantly — an executive layer whose power is structurally capped no matter who occupies it. And the constraint has to run double: thin elites and thin bureaucracy, simultaneously, because both are organizational layers that grow like cancers when unconstrained — and the watchdog layer built to constrain the first is exactly where the surplus of the second goes to lodge.
Gaetano Mosca, the Italian political scientist who founded elite theory, stated the oldest regularity in politics as a law: every society is ruled by an organized minority over an unorganized majority — not sometimes, always — because a minority that acts as one body confronts a majority that acts as millions of separate ones. Robert Michels, testing the law against the most democratic organizations he could find — his own socialist parties — sharpened it into the iron law of oligarchy: whatever the founding ideals, the few who run the machine end up ruling the members. Mancur Olson, again, supplied the economics: organization is a cost, and the arithmetic of who bears it is lopsided — a small group with concentrated stakes finds organizing cheap and profitable, while the vast group with diffuse stakes finds every hour of it a donation to strangers. Concentrated interests organize; diffuse ones stay home. This is the failure mode majority rule's arithmetic pretends away: it assumes the majority can act, and the majority structurally cannot — which is why the lobby beats the public on every issue too small to riot over.
But read the law closely and it is not a law about people — it is a law about the price of coordination. The minority rules because organizing has always been expensive: full-time, professional, worth the cost only to those with concentrated stakes. Every constitution to date has accepted that price and tried to referee its consequences. This proposal's wager is that the price is a technological variable, not a constant — and the design's mechanisms are, one by one, coordination infrastructure for the majority. The open forum makes agenda-setting costless, so surfacing what matters no longer requires a lobby. Delegation is coordination without attendance — millions organizing through granular, revocable trust instead of meetings none of them can afford to hold. Revocation cycles let the majority act in concert without ever assembling; the terminal powers — verdict, removal, the kill — are deliberately the ones that need no assembly at all. And where coordination stays irreducibly expensive — the sustained craft of drafting and passing — the power is held in small tiers that are capped, rotated, refreshed by lot, and revocable (chapter 4). The elites will always coordinate; no design changes that. What this one changes is the monopoly. The iron law priced the majority out of organization; the infrastructure prices it back in.
Run the pattern across history and the lesson sharpens: no governance system has yet avoided all of these failure modes at once, and each great system died of the one it never optimized against. Athens built participation and died of no continuity. Rome built continuity and died of elite overproduction. Feudalism made protection the rent — payment extracted for access, not earned by production — and died of the extraction it never stopped. The long survivors — Venice's millennium, Switzerland's seven centuries, Iceland's three-century Althing (the open-air assembly that governed the island without a king) — were all small, and all unusually good at making their elites replaceable. The design brief writes itself: cover the whole failure table at once, or the uncovered row is where the system dies.
Underneath all of these sits the oldest failure mode: building systems that work when the right person is in charge. Any system whose performance depends on the virtue of its occupant has no error-correction — it is waiting for its variance to kill it. Rome's Five Good Emperors were an adoption-selection fluke that reverted to Commodus the moment heredity reasserted itself. History offers no reliable supply of Trajans — of rulers who hold unchecked power and use it well. The design target is therefore an executive that structurally cannot accumulate excessive power, constrained by the layers beneath it — with the containment running both directions, because the constraint layer itself is where surplus elites lodge and extract.
Each of these failure modes is developed at length in the whitepaper's diagnosis document — elite overproduction and competitive saturation, institutional sclerosis, and epistemic fragmentation, the loss of any shared basis for settling what is true — with the historical evidence behind them. This chapter is the short version; the diagnosis is the long one.