fpf.tools
Library / First Principles Framework (FPF) - Core Conceptual Specification
Jump to passage
In this reading
Source · lines 60388–60475

Link to current text

Published source confirmed at last check

Source changed 2026-10-03 11:52:20 UTC · snapshot created 2026-10-03 11:53:41 UTC · last check 2026-10-03 14:10:16 UTC

C.23:4 - Solution — Method‑SoS‑LOG: deductive shells over Eligibility & Evidence

C.23:4.1 - Objects & heads (LEX/I‑D‑S)

Tech heads; Plain twins are published via UTS. MethodFamily (registered in G.5) carries Eligibility and artefact identity; MaturityCard (this pattern) carries evidence‑aware maturity; SoS‑LOG.Rule (this pattern) is an executable rule schema; evaluating a rule returns one of {Admit | Degrade(mode) | Abstain} for a (TaskSignature, MethodFamily) pair. A qualifying description episteme uses …Description; …Spec names that same episteme only after the E.10.D2 specification-use gate grants the named use.

C.23:4.2 - Rule schema (normative)

For each MethodFamily f, author an executable rule set:

LOG.Deduce_f(TaskSignature S2) → {Admit | Degrade(mode) | Abstain}

with the following branch obligations:

R0 — CG-Spec gate (precondition). For the exact G.5 registry row and MethodFamily, verify the cited CG-Spec.MinimalEvidence and EvidenceProfile for every CHR characteristic used by the family’s acceptance clauses and flows, under the declared claim scope and selected slices, qualification window, and intended selector use. Failure ⇒ Abstain with reasons. Publish the consulted CG-Spec, EvidenceProfile, registry, and policy editions. Rationale: selector legality requires the CG‑Spec gate to be explicit, not implicit in prose. Publish associated ReferencePlane notes alongside the consulted ids.

R0.QD — QD/OEE pre‑gates (if applicable). If S2 declares CharacteristicSpaceRef/ArchiveConfig/EmitterPolicyRef or PortfolioMode=Archive, verify: (i) CharacteristicSpaceRef characteristics are CHR‑typed, d≥2, ReferencePlane per characteristic declared; (ii) ArchiveConfig is lawful (topology, resolution, K>0, InsertionPolicyRef, DistanceDef with edition id and declared metric/pseudometric status); (iii) EmitterPolicyRef present (with edition id); (iv) resolve DominanceRegime; if absent, use default= ParetoOnly. Failure of any ⇒ Abstain with reasons.

R1 — Admit. Admit IFF (a) S2 satisfies Eligibility predicates of f (tri‑state aware), (b) the exact EvidenceProfile minima referenced by Acceptance/Flows for f are met for the declared claim scope and selected slices, qualification window, and intended selector use (post R0), (c) all relevant CAL.AcceptanceClauses (G.4) evaluate to true under lawful CHR comparisons, (d) any maturity gating (e.g., a floor on Maturity rungs) is expressed as an AcceptanceClause and referenced here by id (no acceptance thresholds inside LOG). LOG never sets acceptance thresholds; its rules use and cite Acceptance verdicts.

R2 — Degrade. Apply R0 and, when applicable, R0.QD before any degrade branch, including U2, U3, and R7. Failure returns Abstain for the attempted use. After those gates pass, return Degrade(mode) only through a declared family branch when an admitted unknown or an unmet acceptance condition permits a narrower scope or execution mode under the applicable CAL failure behavior; mode ∈ {scope-narrow | sandbox | probe-only}. An eligibility violation still requires R3 abstention. Record the exact S2 unknowns or unmet conditions, narrowed claim scope or execution mode, qualification window, governing policy edition, and result. If the branch changes the intended use, re-evaluate R0 and eligibility for that bounded use before relying on it. LOG-Degrade never changes CHR scales or planes or turns a failed CAL verdict into a pass. Note (CAL vs LOG). CAL‑level degrade.order (fall‑back to order‑only comparisons) is governed by G.4/CG‑Spec and is not a LOG mode. SoS‑LOG never overrides CAL outcomes; a LOG branch only narrows Scope(G) or execution mode (e.g., sandbox, probe‑only), it does not alter CHR scales or admissible orders. probe‑only MUST cite an E/E‑LOG policy id (exploration budget) and Acceptance‑bound guards.

R3 — Abstain. If S2 violates Eligibility or R0 fails, return Abstain with the failed rule, policy edition, evidence profile, claim scope, qualification window, and reasons. Abstain is mandatory for illegal CHR operations and when a conclusion depends on an F.9 Bridge, kind relation, or plane relation that has not been established.

R4 — Relation and loss routing. Cite an F.9 Bridge, kind relation, or plane relation only when the admission decision actually relies on that obtaining relation. Record its participants, direction, what meaning is preserved and what is lost, receiving use, and applicable policy edition. When the admission use makes a separate named assurance claim, identify its exact target claim and receiving use under B.3. Apply a supported loss penalty only under that assurance policy’s declared rule; route it to R_eff only, leaving F and G unchanged. A changed registry row, evidence profile, claim scope, qualification window, or intended use is not by itself a crossing.

R5 — Proof hooks. Every branch MUST cite Evidence Graph Ref (A.10), the lane tags (TA/VA/LA) and freshness windows required by its cited CG-Spec.MinimalEvidence and EvidenceProfile, and Bridge ids + loss notes when the branch relies on a Bridge; the decision is SCR‑visible. When G.6 EvidenceGraph is present, also publish EvidenceGraph path id(s) for the branch (admit/degrade/abstain). A branch verdict is not its own evidence basis.

R6 — QD archive / PortfolioMode semantics (if applicable). If PortfolioMode=Archive, G.5 selection after Admit may return a QD archive (per ArchiveConfig) instead of only a Pareto set. Unless CAL authorises DominanceRegime=ParetoPlusIllumination (policy‑id recorded in SCR), IlluminationSummary is a report‑only telemetry summary and any coverage/regret are telemetry metrics (reported) that do not affect dominance.

R7 — GeneratorFamily branches (open‑ended). If S2 includes GeneratorIntent, SoS‑LOG MUST: (i) verify EnvironmentValidityRegion is declared and lawful; (ii) verify TransferRulesRef exists; if unknown ⇒ Degrade(scope‑narrow) or Abstain per family policy; (iii) treat the selection surface as pairs {environment, method}; publish coverage/regret and IlluminationSummary as report‑only telemetry (IlluminationSummary = telemetry summary; coverage/regret = telemetry metrics); dominance participation per R6.

R8 — Telemetry & Refresh hooks. On any illumination increase or archive change, publish the current editions and any actual edition increments for CharacteristicSpaceRef/DistanceDefRef/EmitterPolicyRef and the applicable policy‑id (Emitter/Acceptance); expose PathSliceId for refresh/decay in SCR only when an E.18 path slice is current.

Aphorism. “Admit on admissibility and sufficiency; degrade on uncertainty; abstain on inadmissibility.”

C.23:4.3 - Maturity ladder (poset, not a scalar; Description, not Spec)

Publish one editioned MaturityCardDescription for the exact evaluated MethodFamily, G.5 registry edition, evidence profile, claim scope and selected slices, qualification window, and intended admission use (UTS enum ids; scale kind = ordinal; reference plane declared). Do not embed acceptance thresholds here; an admission floor remains a G.4 AcceptanceClause cited by R1.

  • L0 — Anecdotal. Claims exist; lanes sparse; examples ad‑hoc.
  • L1 — Worked‑Examples. Multiple worked examples with lane tags and Scope slices declared; no replication yet.
  • L2 — Replicated. Independent replications identify their distinct bearers or operating conditions and declare the claim scope and selected slices, source and method editions, and qualification windows used; lane separation is observed and decay windows are explicit.
  • L3 — Benchmark‑Severe. Repeated wins or parity on community baselines or severe tests; cross‑Tradition bridges declared with loss notes.

Optional rung (for QD/OEE‑heavy families; ordinal, closed enum):

  • L4 — QD‑Hardened. Archive stability under declared InsertionPolicy/DistanceDef editions; reproducible IlluminationSummary improvements under controlled budgets; OEE generators pass EnvironmentValidityRegion severe tests.

Norms. M1. The ladder is lane‑aware (TA/VA/LA) and freshness‑aware; it is not a global numeric score. Declare Scale kind=ordinal and the closed enumeration of rungs; register the enum at UTS (twin labels; editioned). M2. Transitions MUST be justified by EvidenceGraph paths (once G.6 is available) and published at UTS; missing anchors ⇒ no advance. M3. Any maturity floor used for admission—for example, a run-critical selector use requiring at least L2—MUST be authored as a CAL.AcceptanceClause and cited by R1 with its policy edition, claim scope, qualification window, and verdict; SoS-LOG does not embed acceptance thresholds. M4. Declare the MaturityCard reference plane. If an admission decision relies on a relation to another plane, cite that exact obtaining plane relation, its direction and loss, and the applicable policy edition; a supported loss penalty selected under R4 affects R_eff only.

Rationale note. Treating maturity as a poset aligns with B.3’s requirement for lawful comparisons and avoids scalarisation across ordinal/ratio scales; assurance penalties selected under R4 affect R_eff, never F/G.

C.23:4.4 - Unknowns & Shift classes (tri‑state discipline)

U1. (LEX). Enumerations for Degrade(mode) and Maturity rungs MUST be declared as closed value sets and registered at UTS (twin labels). Lexical SD (E.10) applies. U2. A live S2 characteristic or predicate admits unknown only when its C.22 value rule permits it; unknown MUST map to a branch (Degrade or Abstain) declared on the family (no coercions). Each branch publishes a branch‑id and (where used) a mode from a closed enum registered at UTS (LEX enum clarity). U3. ShiftClass semantics follow C.22. If ShiftClass ∈ {covariate‑shift, concept‑drift, adversarial} or unknown, default outcome is Degrade(scope‑narrow) unless a CAL.AcceptanceClause explicitly guards the regime.

C.23:4.5 - Publication & wiring

W1. For each evaluated MethodFamily, publish an editioned MaturityCardDescription naming the registry edition, evidence profile, claim scope, qualification window, reference plane, and intended admission use; register the SoS-LOG rule ids. RSCR tests cover Admit, Degrade, Abstain, and unknown paths. Relation and loss-policy ids appear only where a branch actually relies on them. W2. Admissibility Ledger. Publish an editioned AdmissibilityLedger: each selector-facing row names the exact MethodFamilyId, G.5 registry edition, RuleId and rule edition, MaturityRung, EvidenceProfile, claim scope, qualification window, BranchIds, AcceptanceClause and policy ids, decision result, evidence paths, DominanceRegime, PortfolioMode, and any obtaining relation and loss-policy ids actually used. UTS registers the row vocabulary; the ledger records the admission result and its basis. W3. Strategy composition. For a selection composition called a strategy, cite its governing G.5 rule and E/E-LOG policy. W4. Selector (G.5) consumes these rules; results appear in the Dispatcher Report with reasons in/out and cited anchors/bridges.