0001 — the central goal and the `admissible` judgment

• Status: exploratory
• Date: 2026-06-07
• Governed-by: ADR 0012

A list of theories can never answer "is that enough?" — every entry invites "why not one more, or one fewer?" The way out is to stop enumerating concerns and find the one relation whose well-definedness forces exactly these theories as its premises. The set then closes by the structure of a single judgment, not by taste.

The central goal

The substrate exists for one thing:

Enable independently authored, future-unknown participants to become safely usable by each other through substrate-visible evidence — without private integration code, central registries, or ambient trust.

A participant written today should be discoverable and usable by one written years later, and vice versa, to the extent the interaction is externalized as values, facts, proofs, descriptors, and resource contracts. This is not magical mutual understanding; it is progressive explicitness. Systems become understandable because they externalize enough of themselves to be checked.

We state this as one goal deliberately. A collection of goals ("be discoverable", "be governable", "be composable") is unsatisfying for exactly the reason that motivates this whole track: nothing about a collection says why those members and not more, fewer, or different ones. We want the semantics to follow from the goal, not to sit beside it.

The goal, restated as a judgment

Everything the substrate does reduces to one move: admit a protected substrate action at a gate such that any rightful challenger can re-check the account for that admission. A fact entering the shared record is one specialization; a read / listen gate-check, a transient emit, a Cell write over a protected current-value surface, or a realization effect are others — the common abstraction is admission at a protected boundary, not record-entry. Write that move as a single relation:

admissible(actor, verb, operand, evidence)

• actor — the principal performing the effect.
• verb — the operation: the base resource vocabulary assert, retract, read, write, emit, listen, plus the candidate realization verbs build / run (under review — 0003; not yet adopted into the closed core set).
• operand — what is acted upon: a value, an artifact, a cell, a wire, a manifest.
• evidence — the account a challenger can re-check.

A theory is then not "a concern we care about." It is an independent premise that must be discharged for admissible to hold. The theories are the premises of one judgment, so the set is closed by the structure of that judgment rather than by enumeration. "Is the set complete?" becomes "what are the independent premises of admissible?" — a question about one relation, answerable by analysis instead of vote.

Telos and mechanism are the same statement from two ends

The goal (the telos) is safe late-bound use. It cashes out to a mechanism: every rightful challenge to an admission must be dischargeable by evidence alone. admissible is that mechanism written down. Each theory exists to make one dimension of challenge checkable — "by what right?", "in what form?", "is that a real, usable thing?", "can it actually be brought into being here?" The goal motivates the judgment; the judgment generates the theories. Neither half is enough alone: the goal without the judgment is a wish-list; the judgment without the goal is an unmotivated relation.

The verbs bound the question

The grounds on which an admission can fail are not unbounded prose — they are bounded by what can go wrong when admitting one verb on one operand. The base operation vocabulary is small and closed (assert, retract, read, write, emit, listen); the realization verbs build, run are candidates under review (0003), not yet part of the closed core. That turns "what are all possible semantics?" into the far more tractable "what independent checks does each verb-on-operand admission require?" 0002 makes that enumeration into the two tests; 0003 walks the matrix and reads off the candidates.

The invariant the whole judgment rides on

admissible is a check, never a search. Finding the evidence may be expensive, heuristic, language-specific, LLM-assisted, distributed, or wrong; verifying it must be small, deterministic, portable, and trusted. This is logos's search/check split, and stele already runs it (admission = check; read = solve + check, the searcher untrusted, the certifying gate trusted). The split is what lets the periphery reason in open-ended, even undecidable ways while the core stays safe — it is presupposed by every theory and is not itself one.

Where this goes

• 0002 turns "independent premise of admissible" into two operational tests — when a theory is required, and when it is irreducible.
• 0003 applies them and reads the candidate theories off the modal structure of the judgment.
• 0004 shows why this framing makes completeness a decidable question at the core and a deliberately open one at the periphery, dissolving the "nothing bounds the set" worry.

Open edges

• The judgment is still prose-shaped. The next real move is to write admissible as an actual logos-checkable relation and see which premises it cannot be defined without (0004 §3).
• evidence is doing a lot of work. Whether it is one uniform account type or a family indexed by verb is unresolved.
• verb includes build / run, which are not resource-primitive verbs but realization verbs. Whether realization deserves verbs in the core vocabulary, or stays an occupant concern, is exactly the 0003 capability/realization question.