Concept Explainer · LiO v0.1

LiO v0.1, explained

A concise reference for readers who want the core mechanism without the formal notation.

What is an indispensable inferential bridge?

A step in a derivation that a downstream claim genuinely depends on — a premise interpreted, mapped, transformed, or generalized on the way to a conclusion. LiO does not decide which steps are indispensable; that judgment is supplied to it as an input.

What is non-compensatory composition, and the weakest-link principle?

The authorization scope of a derivation cannot exceed what every indispensable bridge permits. If one bridge excludes a claim, no permission supplied by any other bridge can restore it — there is no averaging, scoring, or voting. This is deliberately stricter than most policy-combination systems, which often allow strong support in one place to outweigh a restriction elsewhere.

What are Representation, Adjudication, and Composition?

Representation determines what assertion classes and semantic distinctions exist in the universe. Adjudication determines what each bridge permits, excludes, or leaves unresolved. Composition determines what authorization survives once those judgments have been supplied. LiO v0.1 addresses only Composition — the first two remain upstream, unsolved problems that LiO explicitly does not attempt.

What is actually claimed?

  • That given a complete, correctly supplied set of indispensable bridges and their upstream-adjudicated permissions, LiO deterministically computes what a target assertion is authorized to claim, and preserves genuine unresolvedness rather than forcing a premature answer.
  • That a separate AI model, working blind from the frozen specification alone, reproduced LiO's behavior exactly across 19 test cases (7 public, 12 held-out).

What remains unproven?

  • Whether the assertion universe supplied to LiO actually preserves every semantic distinction that matters (“assertion-space adequacy”) — required, not validated.
  • Whether the upstream permission judgments themselves are reproducible and non-circular — required, not validated.
  • Whether the declared set of indispensable bridges is complete and correct — LiO cannot detect a missing dependency.
  • Whether bridge-wise permission sets are expressive enough for cases where permissions are jointly or conditionally structured, rather than independent.

Where can readers go deeper?

The canonical paper record · Research brief · The claim-ledger research map · Evidence · Methods

Suggested citation
Preble, Micheal Charles. “LiO v0.1: Deterministic Permission Composition Across Indispensable Inferential Bridges” Available at SSRN 7424499, 2026. CC BY 4.0.