Skip to content

PR12.11 — Governed Admitted-Policy Requirement Mapping and Coverage Application v1

Purpose

PR12.11 applies one exact already-admitted PR12.7 domain policy to one exact claim and its complete PR12.9/PR12.10 evidence basis without turning semantic requirement coverage into claim truth or a directional ClaimEvaluation.

The critical architectural fact is that the mapping

EvidenceId X semantically covers requirement Y

is not reconstructible from the existing record set, disposition receipt, lineage receipt, or policy specification. It is a human semantic judgment.

Therefore PR12.11 has an explicit governed boundary:

exact admitted policy authority
+ exact PR12.9 disposition coverage
+ exact PR12.10 lineage basis
        ↓
non-authoritative complete mapping proposal
        ↓
explicit HUMAN terminal review
        ↓
runtime-only review admission authority
        ↓
deterministic application receipt

RAW MAPPING PROPOSAL != AUTHORITY
MAPPING JSON != AUTHORITY
RAW HUMAN APPROVE != AUTHORITY
SERIALIZED REVIEW LEDGER != AUTHORITY

Exact upstream basis

A proposal is built only from exact:

  • EpistemicRecordSet;
  • CapabilityClaimId;
  • as_of;
  • fully replay-validated PR12.9 ClaimEvidenceDispositionCoverageReceipt;
  • fully replay-validated PR12.10 ClaimEvidenceLineageDependenceReceipt;
  • PR12.7 registry with current runtime admission authority;
  • EvaluationPolicyRef;
  • exact PR12.6 specification SHA-256;
  • one explicit requirement application entry for every admitted policy requirement.

The selected policy is resolved only through the hardened PR12.7 registry authority path.

RAW POLICY SPEC != ADMITTED POLICY
SERIALIZED REGISTRY != CURRENT RUNTIME REGISTRY AUTHORITY

Exact policy applicability

The admitted policy must match the target claim by exact:

policy.concept_ref == claim.concept_ref
policy.claim_scope == claim.scope

There is no latest-revision matching, capability-id-only matching, tag subset matching, statement heuristic, fuzzy similarity, or scope widening.

Mapping vocabulary

Each admitted policy requirement receives exactly one:

DomainPolicyRequirementApplicationEntry
    requirement_key
    disposition
    evidence_ids
    rationale

The only v1 dispositions are:

COVERED
NOT_COVERED
UNRESOLVED

COVERED

A human proposal says one or more exact PR12.9 candidate evidence values address the exact semantic requirement.

COVERED requires at least one mapped EvidenceId.

NOT_COVERED

A human proposal explicitly says no evidence mapping is admitted for that requirement on this exact basis.

NOT_COVERED requires evidence_ids == ().

UNRESOLVED

The human proposal does not assert either covered or not-covered for the exact requirement.

UNRESOLVED also requires evidence_ids == ().

No missing-entry shorthand exists. Optional requirements also require explicit entries.

MISSING ENTRY != NOT_COVERED
MISSING ENTRY != UNRESOLVED
MISSING ENTRY != OPTIONAL

Coverage is not bearing

PR12.11 preserves the distinction between semantic coverage and directional evidence bearing.

Mapped evidence may have PR12.9 bearing:

SUPPORTS
CONTRADICTS
INDETERMINATE

and still cover a semantic requirement.

COVERED != SUPPORTS
CONTRADICTS CAN COVER A REQUIREMENT
INDETERMINATE CAN COVER A REQUIREMENT

NOT_RELEVANT evidence cannot satisfy a requirement.

PR12.11 never rewrites PR12.9 bearing or reliability.

Reliability and lineage remain orthogonal

PR12.6 contains no reliability threshold, cardinality rule, majority rule, source weighting, or positive independence requirement. PR12.11 therefore invents none.

LOW != AUTOMATIC NON-COVERAGE
HIGH != EXTRA VOTE
UNASSESSED != AUTOMATIC FAILURE

MULTIPLE MAPPED EVIDENCE != STRONGER COVERAGE
MULTIPLE MAPPED EVIDENCE != INDEPENDENT REPLICATION
PROVEN_SHARED_LINEAGE != AUTOMATIC REQUIREMENT FAILURE
UNRESOLVED LINEAGE != PROVEN INDEPENDENCE

PR12.10 remains mandatory so later code cannot silently reintroduce record-count fiction.

Mapping proposal

ClaimDomainPolicyRequirementMappingProposal is immutable serializable audit data binding:

snapshot_sha256
claim_id
subject_ref
concept_ref
claim_scope
as_of
policy_ref
specification_sha256
policy_review_id
policy_review_sha256
policy_admitted_at
disposition_coverage_sha256
lineage_dependence_sha256
requirement_applications

claim_domain_policy_requirement_mapping_proposal_sha256_v1(...) computes a domain-separated digest over the complete canonical proposal.

The digest is integrity/audit data only.

HUMAN review

ClaimPolicyRequirementMappingReview binds one exact proposal digest to:

review_id
claim_id
policy_ref
mapping_proposal_sha256
HUMAN reviewer ref
APPROVE | REJECT
reviewed_at
rationale

Rules:

  • reviewer kind is exactly HUMAN;
  • review binds the exact proposal digest;
  • reviewed_at >= proposal.as_of;
  • reviewed_at >= policy_admitted_at;
  • one valid ledger lineage may contain at most one terminal review for one exact proposal identity;
  • changed proposal content requires a different digest and review;
  • conflicting terminal review fails closed;
  • REJECT never permits application.

Runtime review admission authority

ClaimPolicyRequirementMappingReviewAdmission is runtime-only and cannot be publicly constructed or serialized.

admit_claim_policy_requirement_mapping_review_v1(...) is the explicit transition that issues authority bound to:

  • exact claim;
  • exact policy ref;
  • exact proposal digest;
  • exact review id/digest;
  • predecessor review-ledger digest;
  • exact one-review transition successor digest;
  • exact current review-ledger digest;
  • current process/PID.

Exact replay is ledger-idempotent but issues fresh runtime authority.

A stale admission fails after review-ledger growth.

POSIX fork inheritance fails closed. The child may explicitly replay admission to obtain child-local authority without duplicating the review. Because PR12.7 policy-registry authority is also process-local, a child performing final application must explicitly replay both the policy-admission authority and the PR12.11 mapping-review authority.

No cryptographic human authentication is claimed.

Deterministic application

After exact proposal replay and exact approved review-admission validation, application is deterministic.

ClaimDomainPolicyRequirementApplicationReceipt binds the complete exact basis plus:

mapping_proposal_sha256
mapping_review_id
mapping_review_sha256
requirement_applications
required_requirement_coverage_complete

The only aggregate boolean in PR12.11 is:

required_requirement_coverage_complete = true
iff
EVERY admitted requirement with required_for_sufficiency == true
has disposition == COVERED

Optional requirement status does not affect this boolean.

REQUIRED COVERAGE COMPLETE != CLAIM SUPPORTED
REQUIRED COVERAGE COMPLETE != CLAIM TRUE
REQUIRED COVERAGE COMPLETE != EvaluationConclusion

PR12.12 remains responsible for conservative claim-wide directional evaluation.

Strict replay and serialization

Proposal, review, review ledger, and application receipt have strict schema-v1 dict/JSON serialization. Validation fails closed on unknown/missing fields, duplicate JSON keys, noncanonical encoding, malformed refs/ids/timestamps/enums, duplicate requirement/evidence ids, noncanonical ordering, stale upstream hashes, changed claims/policies, and post-construction semantic corruption.

A JSON-restored final application receipt may validate only after the complete current upstream, policy-authority, proposal, and HUMAN review-admission replay succeeds.

Public API

DomainPolicyRequirementApplicationDisposition
DomainPolicyRequirementApplicationEntry
ClaimDomainPolicyRequirementMappingProposal
ClaimPolicyRequirementMappingReviewId
ClaimPolicyRequirementMappingReviewerKind
ClaimPolicyRequirementMappingReviewerRef
ClaimPolicyRequirementMappingReviewVerdict
ClaimPolicyRequirementMappingReview
ClaimPolicyRequirementMappingReviewLedger
ClaimPolicyRequirementMappingReviewAdmission
ClaimDomainPolicyRequirementApplicationReceipt

build_claim_domain_policy_requirement_mapping_proposal_v1(...)
validate_claim_domain_policy_requirement_mapping_proposal_v1(...)
review_claim_domain_policy_requirement_mapping_proposal_v1(...)
admit_claim_policy_requirement_mapping_review_v1(...)
validate_claim_policy_requirement_mapping_review_admission_v1(...)
require_approved_claim_policy_requirement_mapping_review_v1(...)
apply_admitted_domain_policy_requirements_v1(...)
validate_claim_domain_policy_requirement_application_v1(...)

Plus domain-separated digest and strict serialization helpers for ordinary serializable artifacts.

Non-goals

PR12.11 != AUTOMATIC REQUIREMENT CLASSIFIER
PR12.11 != EvidenceBearing GENERATION
PR12.11 != EvidenceReliability GENERATION
PR12.11 != EVIDENCE DELETION / DEDUPLICATION
PR12.11 != POSITIVE INDEPENDENCE ADMISSION
PR12.11 != REPLICATION COUNT
PR12.11 != MAJORITY VOTE
PR12.11 != EVIDENCE WEIGHTING
PR12.11 != CONFLICT RESOLUTION
PR12.11 != CLAIM-WIDE EvaluationConclusion
PR12.11 != CLAIM TRUTH
PR12.11 != PersonalCapabilityState
PR12.11 != PROGRESSION
PR12.11 != PRESENTATION
PR12.11 != CRYPTOGRAPHIC HUMAN AUTHENTICATION

Intended continuation

PR12.8  complete candidate portfolio                    ✅
PR12.9  complete explicit disposition coverage          ✅
PR12.10 complete deterministic lineage / non-inference  ✅
PR12.11 governed admitted-policy requirement mapping    ← this PR
PR12.12 conservative domain-sufficient evaluation
then     end-to-end capability inference audit/demo