Institutional Legibility Theory

Most institutions are, in a narrow legal sense, already transparent. Their constitutions are public, their cap tables filed, their operating agreements available on request, their bylaws posted somewhere on a website. Almost no one reads any of it. Almost no one could read it usefully even if they did — the documents are long, jargonized, internally cross-referenced, and written for lawyers reviewing other lawyers' work.

We treat this gap — between what is technically disclosed and what is actually understandable — as the central failure of modern institutional life. Disclosure rewards the institution; legibility rewards the participant. A platform that aspires to register institutions honestly must measure itself against the second standard, not the first.

Concretely: a 60-page operating agreement uploaded as a PDF satisfies disclosure. A one-screen structured summary that lets a pledger correctly answer 'who can override the vote I just cast?' satisfies legibility. Only one of these is evidence that the institution can be held to its own rules.

Every governance scandal of the last two decades shares a common shape: an institution described itself, persistently and sincerely, in language that did not match its actual decision rules. The cooperative whose board could not be recalled by members. The DAO whose multisig sat with three co-founders. The non-profit whose 'community council' was advisory only. The platform whose 'user trust and safety board' had no binding authority over the founder.

We do not treat these as isolated failures. We treat them as the predictable output of a system that allows institutions to be evaluated on their adjectives. The register therefore refuses adjectives as evidence. Self-description is recorded, but never displayed as a governance claim. What is displayed are structured fields: who appoints whom, who can override whom, who can be recalled by whom, on what threshold, after what delay, under what jurisdictional authority.

This is not anti-narrative. Initiatives are welcome to write at length about their mission, their history, their values. That writing is theirs. But it is filed under 'mission' and 'narrative', not under 'governance'. Nothing an initiative says about itself is allowed to overwrite a fact about how it decides.

A casual reader needs a five-second answer: can the founder override the members, yes or no. A more engaged reader needs a one-minute summary: governance model, voting rule, amendment rule, founder veto and its scope. A serious reader — a journalist, a regulator, a prospective contributor of meaningful resources — needs the five-minute structural read: the governance map, the appointment chain, the recall thresholds, the history of amendments. A constitutional lawyer needs the deep archive: every diff, every queued change, every appeal.

The register is designed so that all four readers reach the same conclusion about who holds power. They differ only in how much depth they choose to verify. The casual reader is not lied to by simplification; the lawyer is not slowed down by repetition. This is what we mean by layered comprehension, and it is the operating definition of legibility we apply to every screen.

The opposite design — and the dominant one in incumbent governance UI — is layered confusion: a friendly summary that omits the binding rule, an interactive dashboard that does not surface the override clause, a 'transparency report' that excludes the structural levers. We design against this on every page.

Most platforms invert this. Signup is hard — phone verification, captchas, KYC checklists — and governance is trivial. Once you are 'in', the platform makes few further distinctions between you and a founder.

We do the opposite. A pledger expresses a non-binding intent and faces almost no verification, because they hold no power over anyone. An organizer takes on visible representational authority and must clear the next tier. Anyone with binding authority over other members — anyone who can override a vote, fire an officer, amend the charter — must clear the highest tier, including separation between requester and reviewer and a public appeals path.

This rule is normative as well as procedural. It says: the cost of being recognized as an authority should fall on the person seeking the authority, not on the people the authority will be exercised over.

A board adds one observer seat. The next year it makes the observer a voting member. The year after, the threshold for amending the founding documents is quietly lowered from a supermajority to a simple majority. Each change, read alone, is reasonable. Read together over five years, they describe a different institution than the one members originally joined.

We treat this as the most common, and most under-reported, governance failure. The register records every constitutional amendment as a permanent diff, attached to the initiative for its lifetime, with the prior text preserved verbatim. Consequential changes are queued and held for a public delay window before they take effect, so that drift becomes visible while it can still be contested rather than only after it is irreversible.

Drift is also why we keep a 'prior governance model' field next to the current one. A reader should be able to see, at a glance, not only what the institution claims to be today but what it used to claim, and when it changed its claim.

A great deal of confusion in this space comes from platforms that conflate three distinct functions: a place where members can find one another (a registry), a place where members can promise resources (a pledge book), and a place where money actually changes hands (a fundraiser or broker-dealer).

We intentionally only do the first two. The register makes an institution legible; the pledge ledger records that members signaled intent. Anything financial after that point is between the initiative, its members, and the relevant jurisdiction. This is not a limitation we are apologetic about — it is the design. A register that took custody of funds would acquire incentives misaligned with telling the truth about its initiatives.

Governance theater is the production of democratic-feeling rituals — open comment periods, advisory polls, member surveys, 'community calls' — that have no binding effect on the decision actually being made. It is not always cynical. Sometimes it is the residue of good intentions running into a structure that was never amended to honor them.

We design against theater at the field level. A 'community vote' field is meaningless unless paired with a 'vote binding' field. A 'member assembly' field is meaningless unless paired with an 'amendment rule' field. Every free-text 'about us' field is a small failure of this discipline, because it admits as evidence something that is not, by our definition, evidence.

The discipline is not perfect — institutions are messy, and our schema will always lag reality. But the orientation is firm: when in doubt, prefer a structured field. When that is not possible, prefer a structured caveat. Never prefer a paragraph.

The Stewardship Team, Review Circle, and Appeals Circle are public roles with public authority. The platform's own governance map is published at /platform-governance and is evaluated by the same questions we ask of anyone we register: who appoints whom, who can override whom, who can be recalled by whom, on what threshold, after what delay.

This is not symbolic. It is the only honest position. A register that demands structural transparency from its members while refusing it of itself would simply become another institution whose adjective ('neutral', 'community-led', 'independent') was treated as evidence.

If the platform itself drifts, the same machinery applies: amendments are diffed, queued, and held. The register watches the register.