Skip to content

Governance

The current phase — one-person authority, declared

For year one of the public canon — probably year two as well — the arbiter of inclusion is Prayas Abhinav.

This is stated plainly. There is no review board. There is no editorial committee. A single person decides whether a variant passes the four tests.

This arrangement is declared because the alternatives fail in predictable ways:

  • Pretending to a distributed review board that does not yet exist would be exactly the rhetorical separation SDC diagnoses: the language of separation without the structural fact of it.
  • Holding on to single authority in perpetuity would contradict the canon's claim to openness. The canon is meant to grow; an authority that does not grow with it becomes a bottleneck and eventually a legitimacy problem.

The honest middle path is a self-dissolving autocracy with a fixed trigger: declare the current phase, declare what will end it, and structure the process so the transition is mechanical when the trigger fires.

The transition trigger

When five externally authored variants have been admitted to the canon, a review panel forms from the authors of those variants. Prayas becomes one arbiter among several.

This is fixed. It is not subject to reinterpretation. On the day the fifth external variant is admitted:

  • The five authors are invited to form the panel.
  • A decision-making protocol is adopted for multi-arbiter disputes.
  • Prayas steps back from chair to participant; governance becomes genuinely distributed.
  • The succession of authority is recorded publicly.

The planning for this transition begins before the trigger fires. A draft of the panel's constitution should exist by the time the third external variant is admitted, so the transition is ready when it is needed.

What the arbiter does, today

The arbiter has two distinct jobs. They are kept separate because they carry different weight.

Variant admission — editorial work. Reading variant submissions, applying the four validation tests in writing, recording the result on the variant's review page, admitting / returning for revision / declining, and publishing the review record with each admitted variant so the reasoning is visible and contestable. A new variant joins the canonical catalogue only after this review.

Adoption declarations — not editorial work. Adoption declarations are self-declared. The arbiter does not approve them at intake; there is no review queue. A project picks a form of follow, the badge is generated, the embed code is copied, and the declaration is listed. No waiting, no gatekeeping. The four tests on the Derivation protocol page are honesty conditions the declarer holds themselves to — the self-check on the form surfaces them; the declarer is on their honour.

The arbiter retains one residual power over declarations: revocation. A declaration that turns out to be structurally dishonest — alignment claimed where no artefact performs the separation, derivation claimed where no variant is being authored — may be removed from the public list, in writing, with the reasoning recorded. Revocation is rare. Honesty is enforced primarily by the community of practitioners reading each other's claims, not by a single reviewer.

Why self-declaration, like Creative Commons

Anyone can place a Creative Commons licence on a work without asking permission. Creative Commons does not review CC-BY claims at intake. The badge does its work because adoption is frictionless and the community holds adopters to the meaning of the badge they chose to embed. SDC follows the same model: declare the form of follow that fits your artefact, embed the badge, list yourself. If your alignment claim does not survive scrutiny, the community will say so before the arbiter has to.

The alternative — a review queue with one arbiter — would mean low double-digit adoptions per year and a long tail of submissions waiting for attention. That is exactly the friction that kills uptake. The canon's job is to publish a clear principle, a clear protocol, and clear honesty conditions. The declarer's job is to declare honestly. The arbiter's job, on declarations, is to revoke when honesty has clearly broken down.

Variant authoring is different. Adding a new variant to the canon — a derivation that produces a new canonical text — is editorial work and requires review against the four tests. Declaring that one is deriving (in the badge sense) is self-served; getting that derivation admitted to the canonical catalogue at /variants/{slug}/v1/ is reviewed.

What the arbiter does not do

  • Does not certify projects. Canon inclusion and declaration acceptance are not endorsements.
  • Does not adjudicate disputes between adopters. A project that has declared alignment and has its honesty questioned is answerable to the questioner, not to the arbiter.
  • Does not keep reviews private. Every admission, and every rejection, has a written review record.
  • Does not change the derivation protocol by fiat. Changes to the protocol are themselves published at new URLs with their own version numbers. (The protocol you can read today — v1.0 — was drafted as part of the Koher practice's internal canon and made public when the site went live.)

Conflicts of interest

Prayas is both the arbiter and an author of variants. This is an acknowledged conflict. The mitigations:

  1. Internally authored variants (those by Prayas or Koher) are reviewed against the same four tests and the same review record is published.
  2. External reviewers may be invited informally to comment on internally authored variants. Such comments, when offered, are published alongside the review.
  3. The transition trigger above is designed to resolve this conflict structurally. Once the panel forms, Prayas's own variants are reviewed by other arbiters.

Disputes about acceptance

A variant's acceptance may be disputed by any practitioner in the domain. The dispute is submitted in writing and published on the variant's review page. The arbiter responds in writing. If the dispute raises a substantive issue the initial review missed, the variant may be returned to Under Review.

The dispute does not become a negotiation. The four tests are the criteria. A dispute either shows that a test was misapplied, or it does not. The record of the exchange is public.

Changing this governance document

This document is itself versioned. Substantive changes produce a new version at a new URL; past versions remain live. A project that adopted SDC under governance v1 can point at what it adopted under.

Why this structure

The governance structure mirrors the principle the site is built around. Decisions are made by an identifiable arbiter (the judgement work). The reasoning is written down and published (the language work). The two are structurally apart: the verdict on a variant stands on the written review, not on the arbiter's fluency. An adopter can audit the reasoning, check it against the protocol, and dispute it if the protocol was misapplied.

The canon's own governance is, in this narrow sense, SDC about SDC.