# Vercy AI instruction - YAML 1.2 (JSON-compatible) { "vercy": "1.0-draft", "publication": { "status": "published", "adjudicationStatus": "reviewable-draft", "publishableCanonical": false, "generatedAt": "2026-09-02T16:23:15Z", "synthesisSha256": "8749970690d63778aa413333be80667d67f1fb37df6fb0aa09a57fdae00c6a59", "providerMode": "single-provider-waiver", "providers": [ "Claude" ], "waivedProviders": [ "Grok" ] }, "metaModel": { "id": "WM-KNW-007", "registryId": "vr.wm-knw-007", "name": "Claim / Proposition", "version": "0.3.0-research.1", "previousVersions": [], "entryKind": "entity", "family": "World Models", "category": "Information and virtual systems", "industry": [ "Cross-industry" ], "domain": [ "INF.KNW.CLM" ], "tags": [ "claim", "proposition", "inf.knw.clm" ], "status": "published" }, "canonicalUrl": "https://ver.cy/models/wm-knw-007-claim-proposition/", "sourceUrl": "https://github.com/ver-cy/world-models/tree/feat/mega-model-registry/research/runs/wm-knw-007", "model": { "registry_id": "vr.wm-knw-007", "model_id": "WM-KNW-007", "name": "Claim / Proposition", "entry_kind": "entity", "purpose": "Provide a format-neutral record of a truth-apt assertion: what proposition is asserted, under what scope and qualification, by whom and with what standing, how it relates to other claims, which evidence and assessments it references, and how the claim record itself is identified, versioned, validated, restricted and disposed of. The model records assertion, not verdicts.", "scope_statement": "WM-KNW-007 covers the claim record as an object of knowledge management: propositional content and canonical form, identity and equivalence, scope and qualification (subject, place, time, conditions, measurand), stated calibrated uncertainty, the assertion act and asserting authority, provenance and ownership of the record, typed claim-to-claim relations, bound references to evidence and assessments, record lifecycle and versioning, well-formedness validation, sensitivity/access classification and disposition, and interoperability projections. It excludes the evidence items themselves, the appraisal of evidence bodies, the application and evaluation of inference schemes, the execution of fact-checking or verification, policy enforcement, and runtime audit trails owned by other systems.", "in_scope": [ "Canonical truth-apt statement, its language, atomicity and optional normalized subject-property-value or logical form", "Claim identity, aliases, equivalence, restatement and translation relationships", "Scope qualifiers: subject or population range, spatial extent, validity interval, conditions, assumptions, scenarios and known defeaters", "Quantitative content of measurement claims: measurand, value, unit, stated coverage interval", "Calibrated qualification stated on the claim: confidence and likelihood terms, scale reference and traceable-account pointer", "The assertion act: claimant, role, illocutionary commitment, first and subsequent appearances, asserting authority and mandate", "Provenance, attribution, derivation, ownership, stewardship and rights of the claim record", "Typed claim-to-claim relations: entailment, specialization, contradiction, rebuttal, preference, supersession", "Bound references to evidence items, external assessments/reviews and argument structures, carrying binding parameters only", "Record lifecycle states, transitions, versioning, patch rules and externally issued retraction or correction references", "Well-formedness shapes, required-field profiles and quality/testability signals for the record", "Sensitivity classification, access policy expression, retention schedule, tombstones and disposition rules", "Interoperability alignments and lossy-projection declarations across PROV-O, schema.org, VC 2.0, AIF, nanopublication and RDF 1.2" ], "out_of_scope": [ "Evidence items, citations, source descriptions, evidence typology, evidence appraisal and evidence lifecycle, which are owned by WM-KNW-008", "Application and evaluation of inference, conflict and preference rules (AIF RA, CA and PA node semantics) and any defeasible-reasoning engine", "Execution of fact-checking, verification or adjudication workflows, and the rating semantics, reviewer processes and audit trails of assessment systems", "Runtime enforcement or evaluation of access and usage policies, and policy-decision logging", "The real-world entities, events and quantities the claim is about, which belong to subject-domain models", "Documents, publications and media that carry a claim, and publisher retraction workflows executed under NISO CREC", "Cryptographic proof, credential status lists and verification procedures of Verifiable Credentials transport", "Agent belief state, trust scoring of claimants or sources, and reputation computation", "Storage engines, serializations and access interfaces (JSON, YAML, Markdown, Git, MCP, MongoDB are projections)", "Generic knowledge-asset register mechanics inherited from the parent model WM-KNW-001" ], "boundary_notes": [ { "neighbor": "WM-KNW-008 Evidence / Citation", "distinction": "This model carries the evidence reference, the binding role (supports, challenges, qualifies), directionality and a locally stated binding weight. It never carries the evidence item's description, provenance, appraisal grade or lifecycle; those are dereferenced from WM-KNW-008 and are never copied into a claim record.", "source_refs": [ "SRC-005", "SRC-006" ] }, { "neighbor": "Argumentation / inference-scheme model (not yet registered)", "distinction": "A claim corresponds to the AIF I-node level. Applications of rules of inference (RA), conflict (CA) and preference (PA), argument-scheme definitions and any evaluation of argument acceptability belong to an argumentation model. This model records only that a claim occupies a premise or conclusion role in an externally identified argument structure.", "source_refs": [ "SRC-005" ] }, { "neighbor": "Assessment / fact-check record (schema.org ClaimReview)", "distinction": "This model holds the reviewable claim and a reference to reviews of it. The review's rating scale, reviewer identity workflow, publication and audit trail are owned by the reviewing system; a review reference is never read as a truth value on the claim record.", "source_refs": [ "SRC-003", "SRC-002" ] }, { "neighbor": "W3C PROV provenance model", "distinction": "PROV terms are used as alignment for attribution, derivation, revision and invalidation of the claim record. This model does not own generic Activity/Agent modelling, provenance-of-provenance bundles or provenance query semantics.", "source_refs": [ "SRC-001" ] }, { "neighbor": "Verifiable Credentials and assertion transport", "distinction": "A VC claim is a subject-property-value assertion carried inside a credential secured by proof. This model treats the proposition independently of transport: credential validity, status and cryptographic verification are owned by credential models and must not be interpreted as claim truth or claim lifecycle.", "source_refs": [ "SRC-004" ] }, { "neighbor": "Metrology / observation models", "distinction": "A quantitative claim references a measurand, a value with unit and a stated coverage interval. The measurement procedure, instrument, calibration and metrological traceability chain belong to a measurement or observation model.", "source_refs": [ "SRC-011" ] }, { "neighbor": "Access-policy enforcement platforms (ODRL evaluators)", "distinction": "This model expresses sensitivity classes and ODRL policies about claim records. ODRL is a policy expression language that defines no enforcement or evaluation semantics, and none is claimed here; decisioning and its logs belong to the adopting Dimension's platform.", "source_refs": [ "SRC-013" ] }, { "neighbor": "WM-KNW-001 (parent knowledge model)", "distinction": "Generic knowledge-asset identity, register governance, stewardship and licensing are inherited through the CHILD link. This model specializes only what makes a record truth-apt: proposition, qualification, assertion and claim relations.", "source_refs": [ "SRC-001", "SRC-006" ] } ] }, "sources": [ { "id": "SRC-001", "title": "PROV-O: The PROV Ontology", "organization": "World Wide Web Consortium (W3C)", "url": "https://www.w3.org/TR/prov-o/", "version_or_date": "W3C Recommendation, 30 April 2013", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:10:00Z", "relevance": "Defines prov:Entity, prov:Activity, prov:Agent and the properties wasAttributedTo, wasDerivedFrom, wasQuotedFrom, wasRevisionOf, hadPrimarySource, generatedAtTime and invalidatedAtTime, plus the qualified-relation pattern and prov:Bundle. Grounds record provenance, attribution, derivation, revision and invalidation for claim records." }, { "id": "SRC-002", "title": "Claim - Schema.org Type", "organization": "Schema.org community (W3C Schema.org Community Group)", "url": "https://schema.org/Claim", "version_or_date": "Schema.org version 30.0, released 2026-03-19", "source_type": "schema", "primary_source": true, "authority_tier": 2, "accessed_at": "2026-09-02T09:12:00Z", "relevance": "Defines Claim as a factually oriented claim that can be the itemReviewed of a ClaimReview, with text, name, about, sameAs, appearance, firstAppearance and claimInterpreter. Grounds canonical statement text, appearance context and the documented absence of formal claim-to-claim relations." }, { "id": "SRC-003", "title": "ClaimReview - Schema.org Type", "organization": "Schema.org community (W3C Schema.org Community Group)", "url": "https://schema.org/ClaimReview", "version_or_date": "Schema.org version 30.0, released 2026-03-19", "source_type": "schema", "primary_source": true, "authority_tier": 2, "accessed_at": "2026-09-02T09:12:30Z", "relevance": "Defines claimReviewed, itemReviewed, reviewRating, author and datePublished, and locates claimant and claim date on the reviewed item. Grounds the assessment-reference boundary: reviews are referenced, their ratings and workflow are not owned here." }, { "id": "SRC-004", "title": "Verifiable Credentials Data Model v2.0", "organization": "World Wide Web Consortium (W3C)", "url": "https://www.w3.org/TR/vc-data-model-2.0/", "version_or_date": "W3C Recommendation, 15 May 2025", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:14:00Z", "relevance": "Defines a claim as 'an assertion made about a subject' in a subject-property-value structure, with credentialSubject, issuer, validFrom/validUntil, credentialStatus and proof. Grounds normalized claim form and the transport-versus-content boundary." }, { "id": "SRC-005", "title": "The Argument Interchange Format (AIF) Specification", "organization": "ARG-tech, Centre for Argument Technology, University of Dundee", "url": "https://www.arg-tech.org/wp-content/uploads/2011/09/aif-spec.pdf", "version_or_date": "AIF specification document, published 2011 (AIF core ontology originating 2006)", "source_type": "ontology", "primary_source": true, "authority_tier": 2, "accessed_at": "2026-09-02T09:16:00Z", "relevance": "Defines the AIF upper ontology: I-nodes (information/propositions) and S-nodes covering applications of rules of inference (RA), conflict (CA) and preference (PA), with schematic form nodes. Grounds the claim = I-node mapping and the exclusion of inference, conflict and preference rule application." }, { "id": "SRC-006", "title": "Nanopublication Guidelines (nschema)", "organization": "Nanopublication community (Nanopublication/Guidelines repository)", "url": "https://github.com/Nanopublication/Guidelines/blob/master/nschema.html", "version_or_date": "master branch guidelines, retrieved 2026-09-02", "source_type": "first-party-doc", "primary_source": true, "authority_tier": 3, "accessed_at": "2026-09-02T09:18:00Z", "relevance": "Specifies the head / assertion / provenance / publicationInfo named-graph pattern, the hasAssertion, hasProvenance and hasPublicationInfo predicates, and the instruction to limit assertion triples to the minimum so that credit attaches to a precisely defined statement. Grounds atomicity, the separation of assertion from record metadata, and projection guidance." }, { "id": "SRC-007", "title": "RFC 3339: Date and Time on the Internet: Timestamps", "organization": "Internet Engineering Task Force (IETF)", "url": "https://www.rfc-editor.org/rfc/rfc3339", "version_or_date": "RFC 3339, July 2002", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:20:00Z", "relevance": "Requires date-time values with a seconds field and a time-offset of 'Z' or a numeric offset, and distinguishes '-00:00' (offset unknown) from 'Z'/'+00:00'. Grounds the model's timestamp rule and the separation of assertion time from ingestion time." }, { "id": "SRC-008", "title": "Shapes Constraint Language (SHACL)", "organization": "World Wide Web Consortium (W3C)", "url": "https://www.w3.org/TR/shacl/", "version_or_date": "W3C Recommendation, 20 July 2017", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:22:00Z", "relevance": "Defines shapes graph, data graph, focus node, node and property shapes, constraint components, conformance and the severity levels sh:Violation, sh:Warning and sh:Info, and separates shapes from the processor that executes validation. Grounds the well-formedness finding and the validation-engine boundary." }, { "id": "SRC-009", "title": "RDF 1.2 Concepts and Abstract Syntax", "organization": "World Wide Web Consortium (W3C)", "url": "https://www.w3.org/TR/rdf12-concepts/", "version_or_date": "W3C Candidate Recommendation Snapshot, 07 April 2026", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:24:00Z", "relevance": "Introduces triple terms and reifying triples (predicate rdf:reifies) that let a statement be named and described without being asserted, alongside named graphs and IRI-based identity. Grounds statement-level identity projection and the asserted-versus-unasserted distinction." }, { "id": "SRC-010", "title": "Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties", "organization": "Intergovernmental Panel on Climate Change (IPCC)", "url": "https://www.ipcc.ch/site/assets/uploads/2018/05/uncertainty-guidance-note.pdf", "version_or_date": "Mastrandrea et al., 2010; carried forward for AR6", "source_type": "public-authority", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:26:00Z", "relevance": "Defines the evidence (limited/medium/robust) and agreement (low/medium/high) axes, the five-level confidence scale, the likelihood scale with probability ranges, the requirement for traceable accounts, and the instruction to specify variables, region, time period and conditions in a finding. Grounds scope qualification and calibrated confidence." }, { "id": "SRC-011", "title": "JCGM 200:2012 International vocabulary of metrology - Basic and general concepts and associated terms (VIM), 3rd edition", "organization": "Joint Committee for Guides in Metrology (JCGM) / Bureau International des Poids et Mesures (BIPM)", "url": "https://www.bipm.org/en/doi/10.59161/jcgm200-2012", "version_or_date": "JCGM 200:2012, 3rd edition (2008 version with minor corrections)", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:28:00Z", "relevance": "Provides the agreed terms measurand, measurement result, measurement uncertainty, coverage interval, coverage probability and metrological traceability. Grounds the quantitative-content finding and the boundary that measurement procedure and traceability chains belong to a metrology model." }, { "id": "SRC-012", "title": "NISO RP-45-2024, Communication of Retractions, Removals, and Expressions of Concern (CREC)", "organization": "National Information Standards Organization (NISO)", "url": "https://www.niso.org/publications/rp-45-2024-crec", "version_or_date": "Recommended Practice, approved 26 June 2024, published 27 June 2024", "source_type": "standard", "primary_source": true, "authority_tier": 2, "accessed_at": "2026-09-02T09:30:00Z", "relevance": "Defines best practice for creating, transferring and displaying retraction, removal and expression-of-concern metadata, with elements classified Essential, Essential if Available and Recommended, so that readers can identify an item's status. Grounds externally issued status-notice references and the rule that notice execution stays with publishers and aggregators." }, { "id": "SRC-013", "title": "ODRL Information Model 2.2", "organization": "World Wide Web Consortium (W3C)", "url": "https://www.w3.org/TR/odrl-model/", "version_or_date": "W3C Recommendation, 15 February 2018", "source_type": "standard", "primary_source": true, "authority_tier": 1, "accessed_at": "2026-09-02T09:32:00Z", "relevance": "Defines Policy, Rule, Permission, Prohibition, Duty, Action, Asset, Party and Constraint as a policy expression language and states that it does not define enforcement or evaluation mechanisms. Grounds access-policy expression on claim records without granting enforcement semantics." }, { "id": "SRC-014", "title": "FAIR Principles", "organization": "GO FAIR International Support and Coordination Office", "url": "https://www.go-fair.org/fair-principles/", "version_or_date": "Current published principles page, derived from the 2016 Scientific Data paper 'The FAIR Guiding Principles for scientific data management and stewardship'", "source_type": "first-party-doc", "primary_source": true, "authority_tier": 3, "accessed_at": "2026-09-02T09:34:00Z", "relevance": "States F1 (globally unique and persistent identifiers), I3 (qualified references to other metadata) and R1.2 (detailed provenance). Grounds the identity priority, the requirement that claim relations and evidence bindings be qualified rather than bare links, and record provenance expectations." } ], "structure": { "bundles": [ { "id": "proposition-content-and-identity", "name": "Proposition Content and Identity", "description": "What proposition the record asserts, in what form, of what kind, and how that proposition is identified and recognised as the same claim across restatements, translations and systems.", "rationale": "Nothing else in the model is well defined until the asserted proposition is fixed as a single truth-apt statement with a stable identifier. Schema.org locates claim content in text/name/sameAs, VC 2.0 fixes a subject-property-value structure, the nanopublication guidelines require minimal precisely defined assertions, and RDF 1.2 supplies IRI and statement-level identity.", "source_refs": [ "SRC-002", "SRC-004", "SRC-006", "SRC-009", "SRC-014" ], "layers": [ { "id": "content-and-form", "name": "Content and Form", "description": "The canonical statement, its logical shape and its type, including the test that decides whether the content is truth-apt at all.", "source_refs": [ "SRC-002", "SRC-004", "SRC-006" ], "findings": [ { "id": "canonical-statement", "name": "Canonical Statement", "description": "The single declarative sentence the record asserts, held in a canonical surface form with a language tag, an optional normalized structured form for machine comparison, and an atomicity rule of one record to one proposition.", "source_refs": [ "SRC-002", "SRC-004", "SRC-006" ], "questions": [ { "id": "cst-q1", "text": "What single truth-apt proposition does this record assert, expressed as one declarative sentence?", "kind": "definition", "answer_data": [ "Canonical statement text", "Declarative-form check result", "Curator note on wording choices" ] }, { "id": "cst-q2", "text": "Which normalized structured form represents the statement for machine comparison?", "kind": "interoperability", "answer_data": [ "Subject reference", "Property or predicate reference", "Value, object or literal with datatype", "Normalization profile identifier" ] }, { "id": "cst-q3", "text": "Does the statement bundle several separately assertable propositions that must be split into distinct records?", "kind": "constraint", "answer_data": [ "Atomicity verdict", "Enumerated candidate sub-claims", "Split or retain decision with rationale" ] }, { "id": "cst-q4", "text": "Is the canonical text a verbatim quotation of the claimant or a curator paraphrase, and in which language?", "kind": "provenance", "answer_data": [ "Verbatim or paraphrase flag", "BCP 47 language tag", "Locator of the quoted passage if verbatim" ] } ], "data_elements": [ { "id": "cst-d1", "name": "Canonical statement text", "description": "The declarative sentence asserted by this record, stored exactly as curated.", "value_kind": "text", "cardinality": "1", "required": true, "source_refs": [ "SRC-002" ] }, { "id": "cst-d2", "name": "Statement language tag", "description": "BCP 47 language tag of the canonical statement text.", "value_kind": "code", "cardinality": "1", "required": true, "source_refs": [ "SRC-002" ] }, { "id": "cst-d3", "name": "Normalized statement form", "description": "Structured subject-property-value or logical rendering of the statement used for comparison and projection.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-004" ] }, { "id": "cst-d4", "name": "Atomicity verdict", "description": "Whether the statement passed the one-record-one-proposition test.", "value_kind": "boolean", "cardinality": "1", "required": true, "source_refs": [ "SRC-006" ] } ], "artifacts": [], "inline_only_rationale": "The canonical statement and its normalized form are the record's own primary field values, not a separately managed deliverable; the nanopublication guidance to keep assertions minimal implies inline triples rather than a document, and treating the statement as an artifact would create a second identity for the same proposition." }, { "id": "logical-form-and-modality", "name": "Logical Form and Modality", "description": "The structural properties that change what the statement asserts: polarity, quantification over the subject range, alethic or deontic modality, tense, and hedging that weakens the assertion.", "source_refs": [ "SRC-004", "SRC-010", "SRC-005" ], "questions": [ { "id": "lfm-q1", "text": "Is the assertion affirmative or negative, and does negation take wide or narrow scope?", "kind": "definition", "answer_data": [ "Polarity code", "Negation scope note", "Contrast statement that the negation excludes" ] }, { "id": "lfm-q2", "text": "Does the claim quantify universally, existentially or statistically over its subject range?", "kind": "classification", "answer_data": [ "Quantifier code", "Proportion or frequency expression if statistical", "Explicit exception allowance" ] }, { "id": "lfm-q3", "text": "What modality does the statement carry: factual, possible, necessary, obligatory or recommended?", "kind": "constraint", "answer_data": [ "Modality code", "Alethic or deontic distinction", "Source of the modal force" ] }, { "id": "lfm-q4", "text": "Which hedges or qualifiers in the wording weaken the assertion, and are they preserved or moved into structured fields?", "kind": "quality", "answer_data": [ "Hedge inventory", "Preserved-in-text or lifted-to-field decision", "Residual vagueness flag" ] } ], "data_elements": [ { "id": "lfm-d1", "name": "Polarity", "description": "Whether the proposition is asserted affirmatively or negatively.", "value_kind": "code", "cardinality": "1", "required": true, "source_refs": [ "SRC-004" ] }, { "id": "lfm-d2", "name": "Quantification", "description": "Quantifier over the subject range, with proportion where the claim is statistical.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "lfm-d3", "name": "Modality", "description": "Alethic or deontic modal force of the assertion.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "lfm-d4", "name": "Hedge inventory", "description": "Qualifying expressions detected in the statement that weaken its assertive force.", "value_kind": "collection", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] } ], "artifacts": [], "inline_only_rationale": "Polarity, quantification, modality and hedges are coded attributes of the statement itself. Externalizing them into an artifact would separate them from the sentence whose meaning they determine and would allow the two to drift apart under patching." }, { "id": "claim-type-and-truth-aptness", "name": "Claim Type and Truth-Aptness", "description": "The gate test that admits only truth-apt content to the register, and the classification of admitted claims by the kind of proposition asserted, which drives which scope and qualification fields are required.", "source_refs": [ "SRC-002", "SRC-010", "SRC-011" ], "questions": [ { "id": "ctt-q1", "text": "Is this content truth-apt, or is it a question, instruction, performative or bare preference that the register must reject?", "kind": "validation", "answer_data": [ "Truth-aptness verdict", "Rejection reason code where not truth-apt", "Referral target for non-propositional content" ] }, { "id": "ctt-q2", "text": "Which claim type applies: empirical, causal, comparative, predictive, definitional, classificatory, normative or evaluative?", "kind": "classification", "answer_data": [ "Claim type code", "Scheme version in force", "Secondary type where the claim is mixed" ] }, { "id": "ctt-q3", "text": "Which scope, measurement and qualification fields does the assigned claim type make mandatory?", "kind": "requirement", "answer_data": [ "Required-field profile for the type", "Fields currently unpopulated", "Promotion blocker list" ] }, { "id": "ctt-q4", "text": "Under what conditions would this claim count as false, and is it falsifiable in principle?", "kind": "quality", "answer_data": [ "Falsification condition statement", "Falsifiability verdict", "Note where the claim is definitional and not empirically falsifiable" ] } ], "data_elements": [ { "id": "ctt-d1", "name": "Truth-aptness verdict", "description": "Whether the content can bear a truth value and may therefore be registered as a claim.", "value_kind": "boolean", "cardinality": "1", "required": true, "source_refs": [ "SRC-002" ] }, { "id": "ctt-d2", "name": "Claim type code", "description": "Concept from the governed claim-type scheme classifying the kind of proposition.", "value_kind": "code", "cardinality": "1..n", "required": true, "source_refs": [ "SRC-010" ] }, { "id": "ctt-d3", "name": "Falsification condition", "description": "Statement of what observation or argument would render the claim false.", "value_kind": "text", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] } ], "artifacts": [ { "id": "ctt-a1", "name": "Claim Type Scheme", "description": "Governed concept scheme of claim types with definitions, mutual-exclusion rules, and the per-type required-field profile that validation enforces.", "media_or_form": [ "controlled vocabulary / concept scheme", "tabular register", "graph serialization" ], "serial": true, "identity_strategy": "Scheme IRI under the owning package namespace plus stable per-concept notations that survive label changes; releases carry a serial and digest, and retired concepts are marked deprecated with a successor rather than removed.", "source_refs": [ "SRC-002", "SRC-010", "SRC-011" ] } ], "inline_only_rationale": null } ] }, { "id": "identity-and-equivalence", "name": "Identity and Equivalence", "description": "How a claim record is named, how aliases are held, and when two differently worded records are treated as the same proposition.", "source_refs": [ "SRC-009", "SRC-014", "SRC-006", "SRC-002" ], "findings": [ { "id": "claim-identifier", "name": "Claim Identifier", "description": "Assignment and resolution of the identifier that names this claim record, following the model's identity priority, with aliases for identifiers issued elsewhere and content digests kept only as secondary keys.", "source_refs": [ "SRC-009", "SRC-014", "SRC-006" ], "questions": [ { "id": "cid-q1", "text": "Which authoritative master-system identifier, if any, already names this claim, and who issued it?", "kind": "identity", "answer_data": [ "Master identifier value", "Issuing system of record", "Issuance context note" ] }, { "id": "cid-q2", "text": "Which governed global identifier or IRI resolves to this claim record and stays resolvable for its retention period?", "kind": "interoperability", "answer_data": [ "Primary IRI", "Namespace authority", "Resolution commitment and expected lifetime" ] }, { "id": "cid-q3", "text": "Which alternative identifiers name the same claim in other systems, and are they aliases or distinct records?", "kind": "relationship", "answer_data": [ "Alias identifier list with issuing system", "Alias-versus-distinct determination", "Evidence used for the determination" ] }, { "id": "cid-q4", "text": "Is any part of the identifier derived from a date, status or confidence value that would make it unstable?", "kind": "constraint", "answer_data": [ "Identifier composition analysis", "Prohibited-component findings", "Remediation plan for non-conforming identifiers" ] } ], "data_elements": [ { "id": "cid-d1", "name": "Primary claim identifier", "description": "The single identifier of record for this claim, selected by the model's identity priority.", "value_kind": "identifier", "cardinality": "1", "required": true, "source_refs": [ "SRC-014" ] }, { "id": "cid-d2", "name": "Alias identifiers", "description": "Identifiers naming the same claim in other systems, each with its issuing authority.", "value_kind": "identifier", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-009" ] }, { "id": "cid-d3", "name": "Canonical content digest", "description": "Digest over the canonical normalized statement, used only as a secondary equivalence key.", "value_kind": "text", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-006" ] } ], "artifacts": [ { "id": "cid-a1", "name": "Claim Identifier Register", "description": "Register binding primary identifiers to claim records and to alias identifiers, recording the issuing authority, the identity-priority tier applied and the resolution commitment for each namespace.", "media_or_form": [ "tabular register", "resolvable identifier service description", "graph serialization" ], "serial": true, "identity_strategy": "The register itself is named by an IRI under the owning package namespace; entries are keyed by the primary claim identifier and are never reused for a different proposition, with retired entries retained as tombstones.", "source_refs": [ "SRC-009", "SRC-014", "SRC-006" ] } ], "inline_only_rationale": null }, { "id": "equivalence-restatement-and-translation", "name": "Equivalence, Restatement and Translation", "description": "The rules and recorded links that decide when a differently worded, summarised or translated statement is the same proposition, a restatement of it, or a genuinely different claim.", "source_refs": [ "SRC-002", "SRC-005", "SRC-006" ], "questions": [ { "id": "ceq-q1", "text": "Which other claim records state the same proposition in different words, and by what test was sameness established?", "kind": "identity", "answer_data": [ "Equivalent claim identifiers", "Equivalence test applied", "Reviewer and decision note" ] }, { "id": "ceq-q2", "text": "Is this record a translation of another claim, and which record is the authoritative language version?", "kind": "provenance", "answer_data": [ "Source claim identifier", "Source and target language tags", "Authoritative-version designation" ] }, { "id": "ceq-q3", "text": "Does a summarised or headline restatement drop scope qualifiers that the original claim carried?", "kind": "quality", "answer_data": [ "Dropped qualifier inventory", "Scope-drift severity assessment", "Corrective action or warning label" ] }, { "id": "ceq-q4", "text": "Which superficially similar claim was deliberately kept separate, and why is it not the same proposition?", "kind": "exception", "answer_data": [ "Near-miss claim identifier", "Distinguishing feature", "Decision record reference" ] } ], "data_elements": [ { "id": "ceq-d1", "name": "Equivalence link", "description": "Qualified link asserting that another claim record states the same proposition, with the test applied.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-002" ] }, { "id": "ceq-d2", "name": "Restatement link", "description": "Qualified link to a paraphrase, summary or headline rendering of this claim, with direction and fidelity note.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "ceq-d3", "name": "Translation link", "description": "Link to another language version, naming the authoritative version.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-006" ] } ], "artifacts": [], "inline_only_rationale": "Equivalence, restatement and translation are qualified reference edges between existing records. Materialising them as a separate artifact would duplicate identity state that the identifier register already holds and would risk the artifact and the edges disagreeing." } ] } ] }, { "id": "scope-qualification-and-uncertainty", "name": "Scope, Qualification and Stated Uncertainty", "description": "The conditions under which the claim is asserted to hold: over what subjects, where, when, under what assumptions, with what measured content, and with what calibrated confidence the asserter attached to it.", "rationale": "The IPCC guidance requires a finding to specify its variables, region, time period and conditions, and to carry calibrated confidence and likelihood terms backed by a traceable account; JCGM 200 supplies the agreed terms for measurands and stated uncertainty. Without these, a claim is untestable and comparisons between claims are unsound.", "source_refs": [ "SRC-010", "SRC-011", "SRC-002" ], "layers": [ { "id": "scope-qualifiers", "name": "Scope Qualifiers", "description": "Subject range, spatial and temporal extent, and the conditions, assumptions and known defeaters that bound the assertion.", "source_refs": [ "SRC-010", "SRC-002" ], "findings": [ { "id": "subject-and-population-scope", "name": "Subject and Population Scope", "description": "The entities the claim ranges over: the named subject or the population definition with inclusion and exclusion criteria, and the level of aggregation at which the claim is asserted.", "source_refs": [ "SRC-010", "SRC-004", "SRC-002" ], "questions": [ { "id": "sps-q1", "text": "Which entity or population is the claim about, identified rather than merely named in prose?", "kind": "identity", "answer_data": [ "Subject reference to a domain identifier", "Population definition text", "Fallback description where no identifier exists" ] }, { "id": "sps-q2", "text": "Which inclusion and exclusion criteria delimit the population, and what is explicitly outside it?", "kind": "constraint", "answer_data": [ "Inclusion criteria", "Exclusion criteria", "Named out-of-population cases" ] }, { "id": "sps-q3", "text": "At which level of aggregation does the claim hold: individual, group, cohort or population average?", "kind": "classification", "answer_data": [ "Aggregation level code", "Warning where the claim may be read as an individual-level assertion", "Ecological-inference caution note" ] }, { "id": "sps-q4", "text": "Does the claim name an identifiable natural person, and does that trigger sensitivity classification?", "kind": "privacy", "answer_data": [ "Identifiable-person flag", "Named individuals inventory", "Referral to sensitivity classification" ] } ], "data_elements": [ { "id": "sps-d1", "name": "Subject reference", "description": "Identifier of the entity or class the claim is about, resolved against a subject-domain model.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-004" ] }, { "id": "sps-d2", "name": "Population definition", "description": "Inclusion and exclusion criteria delimiting the range of the claim.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "sps-d3", "name": "Aggregation level", "description": "The level at which the claim is asserted to hold.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "sps-d4", "name": "Identifiable person flag", "description": "Whether the statement names or renders identifiable a natural person.", "value_kind": "boolean", "cardinality": "1", "required": true, "source_refs": [ "SRC-002" ] } ], "artifacts": [], "inline_only_rationale": "Subject and population scope are per-claim field values and references to subject-domain identifiers. Publishing them as an artifact would either duplicate the domain model's entity descriptions or create a parallel population registry this model has no mandate to own." }, { "id": "spatial-and-temporal-scope", "name": "Spatial and Temporal Scope", "description": "Where and for what period the claim is asserted to hold, kept strictly separate from when it was asserted and when the record was captured.", "source_refs": [ "SRC-010", "SRC-007", "SRC-001" ], "questions": [ { "id": "sts-q1", "text": "For which geographic or jurisdictional extent is the claim asserted to hold?", "kind": "spatial", "answer_data": [ "Named place or region reference", "Jurisdiction code", "Extent geometry or gazetteer identifier" ] }, { "id": "sts-q2", "text": "Over which validity interval does the claim assert its content to be true?", "kind": "temporal", "answer_data": [ "Validity start instant", "Validity end instant or open-ended marker", "Reference period or baseline where comparative" ] }, { "id": "sts-q3", "text": "How does the validity interval differ from the assertion time and from the record ingestion time?", "kind": "provenance", "answer_data": [ "Assertion timestamp", "Ingestion or observation timestamp", "Explanation of any divergence" ] }, { "id": "sts-q4", "text": "Is the claim asserted as timeless, as time-indexed, or as a forecast about a future interval?", "kind": "classification", "answer_data": [ "Temporal character code", "Forecast horizon where predictive", "Expiry or review-by trigger" ] } ], "data_elements": [ { "id": "sts-d1", "name": "Spatial extent reference", "description": "Place, region or jurisdiction over which the claim is asserted to hold.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "sts-d2", "name": "Validity interval", "description": "Start and end of the period for which the claim's content is asserted true, in RFC 3339 with seconds and offset.", "value_kind": "duration", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-007" ] }, { "id": "sts-d3", "name": "Assertion timestamp", "description": "When the claim was asserted by the claimant, distinct from record capture.", "value_kind": "timestamp", "cardinality": "1", "required": true, "source_refs": [ "SRC-007" ] }, { "id": "sts-d4", "name": "Record ingestion timestamp", "description": "When this register captured or observed the claim.", "value_kind": "timestamp", "cardinality": "1", "required": true, "source_refs": [ "SRC-001" ] } ], "artifacts": [], "inline_only_rationale": "These are timestamp and reference fields on the record. Their governance value comes from being adjacent to the statement they qualify; an artifact would add an indirection that makes the crucial event-time versus ingestion-time distinction easier to lose in projection." }, { "id": "conditions-assumptions-and-defeaters", "name": "Conditions, Assumptions and Known Defeaters", "description": "The scenario, background assumptions, ceteris paribus conditions and explicitly acknowledged defeating circumstances under which the claim is or is not asserted to hold.", "source_refs": [ "SRC-010", "SRC-005" ], "questions": [ { "id": "cad-q1", "text": "Under which scenario, model or set of background assumptions is the claim asserted?", "kind": "constraint", "answer_data": [ "Scenario or pathway identifier", "Background assumption list", "Assumption source reference" ] }, { "id": "cad-q2", "text": "Which circumstances, if they obtained, would defeat the claim without falsifying it outright?", "kind": "exception", "answer_data": [ "Defeater description", "Defeater type (rebutting or undercutting)", "Detection signal for the defeater" ] }, { "id": "cad-q3", "text": "Which conditions are load-bearing such that changing them creates a different claim rather than a variant?", "kind": "decision", "answer_data": [ "Load-bearing condition list", "New-record-versus-variant rule applied", "Decision record reference" ] }, { "id": "cad-q4", "text": "Are any assumptions implicit in the wording and not yet stated in structured fields?", "kind": "quality", "answer_data": [ "Implicit assumption inventory", "Lift-to-field decision", "Residual ambiguity flag" ] } ], "data_elements": [ { "id": "cad-d1", "name": "Scenario reference", "description": "Identifier of the scenario, pathway or model under which the claim is asserted.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "cad-d2", "name": "Assumption statement", "description": "A background condition the claim depends on, stated explicitly.", "value_kind": "text", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "cad-d3", "name": "Known defeater", "description": "An acknowledged circumstance that would defeat the claim, with its defeater type.", "value_kind": "object", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] } ], "artifacts": [], "inline_only_rationale": "Conditions and defeaters qualify one specific claim and are meaningless detached from it. Recording a defeater is a note that a defeating circumstance is acknowledged; evaluating whether it actually defeats the claim belongs to an argumentation model, so no local artifact should imply that adjudication happened here." } ] }, { "id": "quantification-and-uncertainty", "name": "Quantified Content and Calibrated Uncertainty", "description": "Measured content of quantitative claims and the calibrated confidence or likelihood term the asserter attached, with the traceable account that justifies it held elsewhere.", "source_refs": [ "SRC-011", "SRC-010" ], "findings": [ { "id": "quantitative-content-and-measurand", "name": "Quantitative Content and Measurand", "description": "For claims asserting a quantity: the measurand as defined, the value and unit, the stated uncertainty as a coverage interval with its coverage probability, and the comparison basis where the claim is comparative.", "source_refs": [ "SRC-011", "SRC-010" ], "questions": [ { "id": "qcm-q1", "text": "What exactly is the measurand, defined precisely enough that two parties would measure the same thing?", "kind": "measurement", "answer_data": [ "Measurand definition", "Quantity kind reference", "Conditions that must hold for the definition to apply" ] }, { "id": "qcm-q2", "text": "What value and unit does the claim assert, and in which unit system?", "kind": "measurement", "answer_data": [ "Numeric value", "Unit identifier", "Unit system and any conversion applied" ] }, { "id": "qcm-q3", "text": "What stated uncertainty accompanies the value, and at what coverage probability?", "kind": "quality", "answer_data": [ "Coverage interval bounds", "Coverage probability", "Whether the interval is stated by the claimant or derived by the curator" ] }, { "id": "qcm-q4", "text": "Against which baseline, reference value or comparator is a comparative quantitative claim asserted?", "kind": "relationship", "answer_data": [ "Baseline or comparator reference", "Comparison operator", "Reference period or population for the comparator" ] } ], "data_elements": [ { "id": "qcm-d1", "name": "Measurand definition", "description": "The quantity intended to be measured, defined per metrological practice.", "value_kind": "text", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-011" ] }, { "id": "qcm-d2", "name": "Asserted quantity value", "description": "Numeric value with unit asserted by the claim.", "value_kind": "quantity", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-011" ] }, { "id": "qcm-d3", "name": "Stated coverage interval", "description": "Interval attributed to the measurand with its coverage probability, as stated in the claim.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-011" ] }, { "id": "qcm-d4", "name": "Comparator reference", "description": "Baseline or comparator against which a comparative quantitative claim is asserted.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] } ], "artifacts": [], "inline_only_rationale": "Value, unit and stated interval are per-claim fields. The measurement procedure, calibration and metrological traceability chain that would justify an artifact belong to a metrology or observation model, so producing one here would assert ownership of measurement semantics this model explicitly disclaims." }, { "id": "calibrated-confidence-and-likelihood", "name": "Calibrated Confidence and Likelihood", "description": "The calibrated qualification the asserter attached to the claim, expressed with a registered scale term, plus the pointer to the traceable account that justifies it. The appraisal of the underlying evidence body is not held here.", "source_refs": [ "SRC-010", "SRC-006" ], "questions": [ { "id": "ccl-q1", "text": "Which registered scale and term express the confidence attached to this claim, and who set it?", "kind": "measurement", "answer_data": [ "Scale identifier and version", "Confidence term", "Setting agent and role" ] }, { "id": "ccl-q2", "text": "Where the claim is probabilistic, which likelihood term or probability range is asserted?", "kind": "measurement", "answer_data": [ "Likelihood term", "Probability range bounds", "Whether the range is asserted or interpreted from prose" ] }, { "id": "ccl-q3", "text": "Which traceable account documents the reasoning behind the stated confidence, and where does it live?", "kind": "evidence", "answer_data": [ "Traceable-account reference", "Owning system of the account", "Resolution status of the reference" ] }, { "id": "ccl-q4", "text": "Is any evidence-body appraisal grade being carried here that should instead be dereferenced from the evidence model?", "kind": "validation", "answer_data": [ "Detected appraisal fields", "Boundary-violation verdict", "Migration action to WM-KNW-008 reference" ] } ], "data_elements": [ { "id": "ccl-d1", "name": "Confidence term", "description": "Calibrated confidence term drawn from the registered scale in force.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "ccl-d2", "name": "Likelihood term or range", "description": "Calibrated likelihood term with its probability range where the claim is probabilistic.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "ccl-d3", "name": "Traceable account reference", "description": "Pointer to the documented reasoning that justifies the stated confidence.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "ccl-d4", "name": "Confidence scale binding", "description": "Identifier and version of the calibrated scale under which the terms are interpreted.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-010" ] } ], "artifacts": [ { "id": "ccl-a1", "name": "Calibrated Uncertainty Scale Registry", "description": "Registry of the confidence and likelihood scales the owning package permits, each with its ordered terms, the probability ranges bound to likelihood terms, and the rule that terms are never compared across scale versions without explicit mapping.", "media_or_form": [ "controlled vocabulary / concept scheme", "tabular register", "machine-readable scale definition" ], "serial": true, "identity_strategy": "Registry IRI under the owning package namespace with per-scale identifiers and per-term notations; each release carries a serial and digest, and records cite scale identifier plus version so that a later scale revision never silently reinterprets an existing claim.", "source_refs": [ "SRC-010" ] } ], "inline_only_rationale": null } ] } ] }, { "id": "assertion-provenance-and-ownership", "name": "Assertion, Provenance and Ownership", "description": "Who asserted the claim and with what standing, how the claim record came to exist, and who is accountable for it.", "rationale": "PROV-O supplies attribution, derivation, primary-source and revision terms; the nanopublication pattern separates assertion metadata from record metadata; VC 2.0 fixes the issuer as the party asserting claims; FAIR R1.2 requires detailed provenance. Together these ground both the assertion act and the record's own lineage and stewardship.", "source_refs": [ "SRC-001", "SRC-004", "SRC-006", "SRC-014", "SRC-002" ], "layers": [ { "id": "assertion-acts", "name": "Assertion Acts and Standing", "description": "The act of asserting: claimant, commitment, appearance context, and whether the claimant held authority to assert.", "source_refs": [ "SRC-002", "SRC-003", "SRC-004" ], "findings": [ { "id": "claimant-and-assertion-act", "name": "Claimant and Assertion Act", "description": "Who asserted the proposition, in what capacity and with what degree of commitment, together with the first and subsequent contexts in which the assertion appeared.", "source_refs": [ "SRC-002", "SRC-003", "SRC-004", "SRC-007" ], "questions": [ { "id": "caa-q1", "text": "Which agent asserted this proposition, identified rather than described by name alone?", "kind": "provenance", "answer_data": [ "Claimant identifier", "Claimant type (person, organisation, automated system)", "Fallback claimant-unknown code with reason" ] }, { "id": "caa-q2", "text": "In what capacity did the claimant assert it: personally, on behalf of an organisation, or as a reported position?", "kind": "authority", "answer_data": [ "Capacity code", "Represented organisation reference", "Reporting chain where the assertion is second-hand" ] }, { "id": "caa-q3", "text": "How strong was the commitment: asserted outright, hypothesised, attributed to others, or quoted without endorsement?", "kind": "classification", "answer_data": [ "Commitment level code", "Endorsement flag", "Wording that signals the commitment level" ] }, { "id": "caa-q4", "text": "Where did the claim first appear and where else has it appeared since?", "kind": "event", "answer_data": [ "First-appearance locator with timestamp", "Subsequent appearance locators", "Medium and audience for each appearance" ] } ], "data_elements": [ { "id": "caa-d1", "name": "Claimant reference", "description": "Identifier of the agent that asserted the proposition.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-004" ] }, { "id": "caa-d2", "name": "Assertion capacity", "description": "The capacity in which the claimant asserted, including representation of an organisation.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-002" ] }, { "id": "caa-d3", "name": "Commitment level", "description": "Degree of assertive commitment, from outright assertion to unendorsed quotation.", "value_kind": "code", "cardinality": "1", "required": true, "source_refs": [ "SRC-002" ] }, { "id": "caa-d4", "name": "Appearance locator", "description": "Locator of an occurrence of the claim, with its timestamp, marking which is the first appearance.", "value_kind": "object", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-003" ] } ], "artifacts": [], "inline_only_rationale": "The assertion act is captured as references to agents and to appearance locations in documents owned by content and publication models. Creating an artifact here would duplicate those documents and would blur the line between recording that an assertion occurred and republishing the material that carried it." }, { "id": "asserting-authority-and-mandate", "name": "Asserting Authority and Mandate", "description": "Whether the claimant held a recognised mandate to assert within a jurisdiction or domain, and the standing that mandate gives the claim, recorded without treating standing as truth.", "source_refs": [ "SRC-004", "SRC-002", "SRC-010" ], "questions": [ { "id": "aam-q1", "text": "Under which mandate, remit or delegated authority did the claimant assert this proposition?", "kind": "authority", "answer_data": [ "Mandate reference", "Granting body", "Mandate scope and limits" ] }, { "id": "aam-q2", "text": "Within which jurisdiction or domain is that authority recognised, and where is it not?", "kind": "spatial", "answer_data": [ "Jurisdiction or domain codes", "Non-recognition notes", "Effective period of recognition" ] }, { "id": "aam-q3", "text": "Is the claim official, advisory or unofficial in standing, and what follows from that classification?", "kind": "classification", "answer_data": [ "Standing code", "Consequences declared by the owning package", "Downgrade triggers" ] }, { "id": "aam-q4", "text": "What prevents authority standing from being read as a truth verdict on the claim?", "kind": "constraint", "answer_data": [ "Separation rule text", "Projection guard for consumers", "Non-conforming usage examples" ] } ], "data_elements": [ { "id": "aam-d1", "name": "Mandate reference", "description": "Identifier of the mandate or remit under which the claimant asserted.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-004" ] }, { "id": "aam-d2", "name": "Recognised jurisdiction", "description": "Jurisdictions or domains in which the asserting authority is recognised.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "aam-d3", "name": "Standing classification", "description": "Official, advisory or unofficial standing of the assertion.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-002" ] } ], "artifacts": [], "inline_only_rationale": "Mandates are instruments issued and maintained by the granting bodies; this model holds a reference and a standing classification only. Producing a local artifact would imply this register certifies authority, which it does not, and would create a competing copy of an external legal or institutional instrument." } ] }, { "id": "record-provenance", "name": "Record Provenance, Ownership and Rights", "description": "How the claim record itself was produced and derived, who owns and stewards it, and under what rights it may be reused.", "source_refs": [ "SRC-001", "SRC-006", "SRC-014" ], "findings": [ { "id": "record-derivation-and-attribution", "name": "Record Derivation and Attribution", "description": "The lineage of the claim record: which source it was derived or quoted from, which primary source stands behind it, which agent or process extracted it, and which revision it succeeds.", "source_refs": [ "SRC-001", "SRC-006", "SRC-007" ], "questions": [ { "id": "rda-q1", "text": "From which prior entity was this claim record derived, quoted or revised?", "kind": "provenance", "answer_data": [ "Derivation source identifier", "Derivation type (derived, quoted, revised)", "Primary source reference where the chain is known" ] }, { "id": "rda-q2", "text": "Which agent or automated process generated the record, and was extraction human, assisted or fully automated?", "kind": "process", "answer_data": [ "Generating agent identifier", "Extraction method code", "Model or tool identifier and version where automated" ] }, { "id": "rda-q3", "text": "Which activity produced the record, and when did it start and complete?", "kind": "temporal", "answer_data": [ "Activity identifier", "Start and end timestamps in RFC 3339 with offset", "Relationship to the separately recorded assertion time" ] }, { "id": "rda-q4", "text": "What confidence attaches to the extraction itself, as distinct from the claim's own stated confidence?", "kind": "quality", "answer_data": [ "Extraction confidence value or code", "Human review status", "Known extraction failure modes for this source type" ] } ], "data_elements": [ { "id": "rda-d1", "name": "Derivation reference", "description": "Qualified link to the entity this record was derived from, quoted from or revised from.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "rda-d2", "name": "Generating agent", "description": "Agent or automated process responsible for generating the record.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "rda-d3", "name": "Extraction method", "description": "How the claim was captured: manual curation, assisted extraction or automated extraction, with tool version.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-006" ] }, { "id": "rda-d4", "name": "Extraction confidence", "description": "Confidence in the fidelity of extraction, kept distinct from the claim's stated confidence.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-006" ] } ], "artifacts": [ { "id": "rda-a1", "name": "Claim Record Provenance Bundle", "description": "A named provenance description set for one or more claim records, holding derivation, attribution, generation activity and revision statements in a form that can itself be attributed and dated, following the PROV bundle and nanopublication provenance-graph pattern.", "media_or_form": [ "named provenance graph", "structured provenance record", "graph serialization" ], "serial": true, "identity_strategy": "Bundle IRI under the owning package namespace, bound to the primary claim identifiers it describes; each bundle carries its own attribution and generation timestamp so that provenance of the provenance is expressible, and bundles are appended rather than edited.", "source_refs": [ "SRC-001", "SRC-006" ] } ], "inline_only_rationale": null }, { "id": "ownership-stewardship-and-rights", "name": "Ownership, Stewardship and Rights", "description": "Which package owns the claim record, who stewards it, which roles may change it, and under what licence or rights the statement and the record may be reused.", "source_refs": [ "SRC-014", "SRC-001", "SRC-013" ], "questions": [ { "id": "osr-q1", "text": "Which package owns this claim record and which named steward is accountable for it?", "kind": "ownership", "answer_data": [ "Owning package identifier", "Accountable steward identity", "Effective date of the assignment" ] }, { "id": "osr-q2", "text": "Which roles may change the statement, the scope qualifiers, the status and the sensitivity class?", "kind": "access", "answer_data": [ "Role-to-field write matrix", "Escalation path for restricted fields", "Segregation-of-duty constraints" ] }, { "id": "osr-q3", "text": "Under which licence or rights statement may the claim record and its canonical text be reused?", "kind": "ownership", "answer_data": [ "Licence identifier", "Rights holder for quoted text", "Attribution requirement" ] }, { "id": "osr-q4", "text": "What happens to ownership when the owning package is retired or transferred?", "kind": "lifecycle", "answer_data": [ "Succession rule", "Transfer record reference", "Interim custodian designation" ] } ], "data_elements": [ { "id": "osr-d1", "name": "Owning package reference", "description": "The package accountable for the claim record.", "value_kind": "reference", "cardinality": "1", "required": true, "source_refs": [ "SRC-014" ] }, { "id": "osr-d2", "name": "Accountable steward", "description": "Named individual or role accountable for the record's content and status.", "value_kind": "reference", "cardinality": "1", "required": true, "source_refs": [ "SRC-014" ] }, { "id": "osr-d3", "name": "Rights statement", "description": "Licence or rights terms governing reuse of the record and its canonical text.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-013" ] } ], "artifacts": [ { "id": "osr-a1", "name": "Stewardship and Rights Declaration", "description": "Declaration naming the owning package, the accountable steward, the role-to-field write matrix, the licence applying to records and to quoted statement text, and the succession rule if ownership is transferred.", "media_or_form": [ "governance declaration document", "structured rights record", "tabular role matrix" ], "serial": true, "identity_strategy": "Declaration IRI under the owning package namespace with a serial per release and a digest; each claim record cites the serial in force at its last status transition so that historical governance terms remain reconstructable.", "source_refs": [ "SRC-014", "SRC-013" ] } ], "inline_only_rationale": null } ] } ] }, { "id": "claim-relations-and-external-bindings", "name": "Claim Relations and External Bindings", "description": "How claims stand to one another and how a claim binds to evidence, assessments and argument structures owned elsewhere.", "rationale": "AIF supplies the typology of inference, conflict and preference relations at the argument level and locates propositions as I-nodes; PROV supplies revision and invalidation; FAIR I3 requires qualified rather than bare references; schema.org documents that it defines no claim-to-claim relations, which makes this a real modelling need. Bindings must carry parameters only, since evidence, review and argument content are owned by other models.", "source_refs": [ "SRC-005", "SRC-001", "SRC-002", "SRC-003", "SRC-014" ], "layers": [ { "id": "typed-claim-relations", "name": "Typed Claim-to-Claim Relations", "description": "Recorded relations between claim records: entailment and specialization, conflict, and preference or supersession, registered without being adjudicated.", "source_refs": [ "SRC-005", "SRC-001", "SRC-002" ], "findings": [ { "id": "entailment-and-specialization", "name": "Entailment and Specialization", "description": "Relations where one claim's content stands to another as implication, generalization, specialization or presupposition, recorded as qualified edges with the ground on which the relation is asserted.", "source_refs": [ "SRC-005", "SRC-014" ], "questions": [ { "id": "ens-q1", "text": "Which other claims does this claim entail, and on what ground is entailment asserted?", "kind": "relationship", "answer_data": [ "Entailed claim identifiers", "Ground code (logical, definitional, domain rule)", "Asserting agent for the relation" ] }, { "id": "ens-q2", "text": "Which broader claim does this one specialize, and which narrower claims specialize it?", "kind": "composition", "answer_data": [ "Broader claim identifier", "Narrower claim identifiers", "Dimension of specialization (scope, population, period)" ] }, { "id": "ens-q3", "text": "What does this claim presuppose that must hold for it to be evaluable at all?", "kind": "constraint", "answer_data": [ "Presupposed claim identifiers", "Consequence if the presupposition fails", "Detection status of failure" ] }, { "id": "ens-q4", "text": "Which relation assertions are curator judgements rather than derived from a formal rule?", "kind": "quality", "answer_data": [ "Judgement-versus-derived flag per edge", "Curator identity", "Review status of contested edges" ] } ], "data_elements": [ { "id": "ens-d1", "name": "Entailment edge", "description": "Qualified edge asserting that this claim entails a target claim, with its ground.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "ens-d2", "name": "Specialization edge", "description": "Qualified edge to a broader or narrower claim, naming the dimension of specialization.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "ens-d3", "name": "Presupposition edge", "description": "Qualified edge to a claim that must hold for this claim to be evaluable.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] } ], "artifacts": [], "inline_only_rationale": "These are typed edges between records in the same register, matching the model's declared TYPED-EDGES link type. Externalising them would duplicate graph state and would tempt consumers to read a curated edge list as a derived inference result, which this model does not produce." }, { "id": "conflict-and-contradiction", "name": "Conflict and Contradiction", "description": "Registration that two claims stand in conflict, with the conflict type and the point of disagreement recorded, and with resolution explicitly not performed here.", "source_refs": [ "SRC-005", "SRC-002" ], "questions": [ { "id": "cfc-q1", "text": "Which claims contradict this one, and is the conflict rebutting the content or undercutting its support?", "kind": "relationship", "answer_data": [ "Conflicting claim identifiers", "Conflict type code", "Point of disagreement statement" ] }, { "id": "cfc-q2", "text": "Is the apparent conflict genuine, or does it dissolve once scope qualifiers are compared?", "kind": "validation", "answer_data": [ "Scope comparison result", "Genuine-or-apparent verdict", "Differing qualifier inventory" ] }, { "id": "cfc-q3", "text": "Who registered the conflict and when, and what status does registration confer on either claim?", "kind": "event", "answer_data": [ "Registering agent", "Registration timestamp", "Status effect declared by the owning package" ] }, { "id": "cfc-q4", "text": "Which system, outside this model, is responsible for resolving or adjudicating the conflict?", "kind": "authority", "answer_data": [ "Resolving system or body reference", "Referral record", "Expected form of the resolution output" ] } ], "data_elements": [ { "id": "cfc-d1", "name": "Conflict edge", "description": "Qualified edge registering conflict with another claim, with conflict type and disagreement point.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "cfc-d2", "name": "Conflict registration event", "description": "Who registered the conflict and when, in RFC 3339 with offset.", "value_kind": "object", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "cfc-d3", "name": "Resolution referral", "description": "Reference to the external body or system responsible for adjudicating the conflict.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-002" ] } ], "artifacts": [], "inline_only_rationale": "Conflicts are registered as qualified edges plus a registration event. A conflict register artifact would read as an adjudication docket and would encroach on the argumentation model that owns conflict-rule application and on the assessment systems that issue verdicts." }, { "id": "preference-and-supersession", "name": "Preference and Supersession", "description": "Recorded orderings between claims: an asserted preference between competing claims, and the supersession chain by which a later claim replaces an earlier one.", "source_refs": [ "SRC-005", "SRC-001", "SRC-012" ], "questions": [ { "id": "pss-q1", "text": "Which competing claim is preferred over which, and by whose criterion?", "kind": "decision", "answer_data": [ "Preferred and dispreferred claim identifiers", "Preference criterion", "Asserting agent and date of the preference" ] }, { "id": "pss-q2", "text": "Which claim does this one supersede, and which claim supersedes it?", "kind": "lifecycle", "answer_data": [ "Superseded claim identifier", "Superseding claim identifier", "Effective instant of supersession" ] }, { "id": "pss-q3", "text": "Does supersession change the earlier claim's status, or leave it valid for its original period?", "kind": "state", "answer_data": [ "Status effect rule", "Earlier claim's retained validity interval", "Consumer guidance for historical queries" ] }, { "id": "pss-q4", "text": "How are supersession chains kept acyclic and traversable to the current claim?", "kind": "constraint", "answer_data": [ "Cycle-detection rule", "Chain-head resolution procedure", "Broken-chain remediation" ] } ], "data_elements": [ { "id": "pss-d1", "name": "Preference edge", "description": "Qualified edge asserting preference between two competing claims, with criterion and asserter.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "pss-d2", "name": "Supersedes reference", "description": "Identifier of the claim this record supersedes.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "pss-d3", "name": "Superseded-by reference", "description": "Identifier of the claim that supersedes this record.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-001" ] } ], "artifacts": [], "inline_only_rationale": "Preference and supersession are directional edges evaluated at query time to find the current claim. Materialising a chain artifact would freeze a view that must stay derivable from the edges, and preference application in argument evaluation is owned by the argumentation model." } ] }, { "id": "external-bindings", "name": "External Bindings", "description": "Qualified references from a claim to evidence items, external assessments and argument structures, carrying binding parameters only.", "source_refs": [ "SRC-005", "SRC-003", "SRC-006", "SRC-014" ], "findings": [ { "id": "evidence-link-binding", "name": "Evidence Link Binding", "description": "The qualified reference from this claim to an evidence item in WM-KNW-008, carrying the target identifier, the binding role, directionality and a locally stated binding weight, and nothing describing the evidence itself.", "source_refs": [ "SRC-005", "SRC-006", "SRC-014" ], "questions": [ { "id": "elb-q1", "text": "Which evidence items are bound to this claim, identified by resolvable WM-KNW-008 identifiers?", "kind": "evidence", "answer_data": [ "Evidence item identifiers", "Resolution status of each identifier", "Binding creation timestamp" ] }, { "id": "elb-q2", "text": "In what role does each bound item stand: supporting, challenging, qualifying or merely cited?", "kind": "relationship", "answer_data": [ "Binding role code", "Directionality", "Curator note on the role assignment" ] }, { "id": "elb-q3", "text": "What weight or relevance does this claim's curator attach to the binding, and how is that distinguished from the evidence's own appraisal?", "kind": "quality", "answer_data": [ "Locally stated binding weight", "Weighting scheme reference", "Explicit statement that evidence appraisal grades are not stored here" ] }, { "id": "elb-q4", "text": "What happens to the binding when the referenced evidence record is revised, withdrawn or deleted by its owner?", "kind": "exception", "answer_data": [ "Dangling-reference handling rule", "Notification expectation from WM-KNW-008", "Local status effect on the claim" ] } ], "data_elements": [ { "id": "elb-d1", "name": "Evidence reference", "description": "Identifier of an evidence item in WM-KNW-008 bound to this claim.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-014" ] }, { "id": "elb-d2", "name": "Binding role", "description": "Role the referenced item plays relative to this claim: supports, challenges, qualifies or cites.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "elb-d3", "name": "Locally stated binding weight", "description": "Curator-stated relevance or strength of this specific binding, explicitly not an evidence appraisal grade.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-006" ] }, { "id": "elb-d4", "name": "Binding status", "description": "Whether the binding is active, dangling or retired following changes at the target.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-001" ] } ], "artifacts": [], "inline_only_rationale": "The binding is a qualified edge with a small parameter set, and by the REFERENCE contract with WM-KNW-008 nothing describing the evidence may be held locally. Any artifact here would inevitably accrete evidence metadata and would duplicate the target model's records." }, { "id": "assessment-and-review-reference", "name": "Assessment and Review Reference", "description": "References from this claim to externally issued assessments, fact-checks or reviews of it, recorded so that consumers can find them without this model owning ratings, reviewer workflow or review audit trails.", "source_refs": [ "SRC-003", "SRC-002", "SRC-012" ], "questions": [ { "id": "arr-q1", "text": "Which external reviews or assessments name this claim as the item reviewed?", "kind": "relationship", "answer_data": [ "Review identifiers and URLs", "Reviewing organisation", "Review publication timestamp" ] }, { "id": "arr-q2", "text": "Which rating scale did each review use, and is that scale interpretable outside its issuing system?", "kind": "interoperability", "answer_data": [ "Rating scale reference", "Scale interpretability note", "Prohibition on normalising ratings across issuers" ] }, { "id": "arr-q3", "text": "What effect, if any, does a review reference have on this claim record's own status?", "kind": "state", "answer_data": [ "Status effect rule", "Cases where a review triggers a dispute status", "Cases where it triggers nothing" ] }, { "id": "arr-q4", "text": "Who owns the review's workflow, evidence and audit trail, and what may this model never copy from it?", "kind": "authority", "answer_data": [ "Owning system reference", "Prohibited copy list", "Escalation route for review corrections" ] } ], "data_elements": [ { "id": "arr-d1", "name": "Assessment reference", "description": "Identifier or locator of an external review or assessment of this claim.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-003" ] }, { "id": "arr-d2", "name": "Reviewing organisation", "description": "Organisation or person that issued the referenced assessment.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-003" ] }, { "id": "arr-d3", "name": "Rating scale reference", "description": "Pointer to the issuing system's rating scale, held as an opaque reference and not normalised locally.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-003" ] } ], "artifacts": [], "inline_only_rationale": "Reviews are documents produced and governed by fact-checking and assessment systems. This model stores only resolvable references, because producing a local artifact would amount to republishing verdicts and would assert ownership of rating semantics and reviewer audit trails that belong to those systems." }, { "id": "argument-structure-role-reference", "name": "Argument Structure Role Reference", "description": "The record that this claim occupies a premise or conclusion role in an externally identified argument structure, mapped to the AIF I-node level, without any local representation of inference-scheme application.", "source_refs": [ "SRC-005", "SRC-002" ], "questions": [ { "id": "asr-q1", "text": "In which externally identified argument structures does this claim appear, and in what role?", "kind": "composition", "answer_data": [ "Argument structure identifiers", "Role code (premise, conclusion, exception)", "Owning argumentation system reference" ] }, { "id": "asr-q2", "text": "How does the claim map to the AIF I-node level, and which node types are deliberately not represented here?", "kind": "interoperability", "answer_data": [ "I-node mapping statement", "Excluded node types (RA, CA, PA, F)", "Crosswalk edition reference" ] }, { "id": "asr-q3", "text": "What prevents an argument role reference from being read as an evaluated inference or an accepted conclusion?", "kind": "constraint", "answer_data": [ "Separation rule text", "Consumer guard in projections", "Examples of prohibited inference" ] }, { "id": "asr-q4", "text": "Which registered model will own argument-scheme definition and evaluation once it exists?", "kind": "decision", "answer_data": [ "Declared gap status", "Candidate owner description", "Interim handling rule" ] } ], "data_elements": [ { "id": "asr-d1", "name": "Argument structure reference", "description": "Identifier of an external argument structure in which this claim participates.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "asr-d2", "name": "Argument role", "description": "Role the claim plays in the referenced structure.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-005" ] }, { "id": "asr-d3", "name": "Node-type mapping note", "description": "Statement of the AIF node level this record maps to and the node types excluded from local representation.", "value_kind": "text", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-005" ] } ], "artifacts": [], "inline_only_rationale": "Only a role reference is held, because AIF rule-application nodes and scheme definitions belong to an argumentation model that is not yet registered. Producing an argument artifact here would create exactly the ownership overlap the boundary review is meant to prevent." } ] } ] }, { "id": "lifecycle-validation-and-record-governance", "name": "Lifecycle, Validation and Record Governance", "description": "The states a claim record passes through, how it is versioned and corrected, how its well-formedness and quality are checked, and how it is restricted, retained and disposed of.", "rationale": "PROV supplies revision and invalidation semantics, NISO CREC supplies the metadata practice for externally issued retractions and expressions of concern, SHACL supplies shapes and severity-graded conformance separated from the processor, and ODRL supplies policy expression that explicitly excludes enforcement. These fix a governance surface for the record without claiming execution.", "source_refs": [ "SRC-001", "SRC-012", "SRC-008", "SRC-013", "SRC-007" ], "layers": [ { "id": "record-lifecycle", "name": "Record Lifecycle and Change Control", "description": "States and transitions of the claim record, the rule separating a new version from a new claim, and references to externally issued status notices.", "source_refs": [ "SRC-001", "SRC-012", "SRC-007" ], "findings": [ { "id": "status-and-state-transitions", "name": "Status and State Transitions", "description": "The permitted states of a claim record, the transitions between them, who may perform each transition, and the terminal timestamps and successor references that transitions set.", "source_refs": [ "SRC-001", "SRC-007", "SRC-012" ], "questions": [ { "id": "sst-q1", "text": "Which status does the record currently hold, and since which instant?", "kind": "state", "answer_data": [ "Current status code", "Status effective timestamp in RFC 3339 with offset", "Prior status" ] }, { "id": "sst-q2", "text": "Which transitions are permitted from the current status, and which roles may perform them?", "kind": "lifecycle", "answer_data": [ "Permitted transition list", "Role authorisation per transition", "Blocked transitions with reasons" ] }, { "id": "sst-q3", "text": "Which preconditions must be satisfied before a record may leave draft and be treated as asserted?", "kind": "requirement", "answer_data": [ "Promotion precondition list", "Outstanding blocker inventory", "Conformance report reference" ] }, { "id": "sst-q4", "text": "What does a terminal status set on the record, and what must remain readable afterwards?", "kind": "retention", "answer_data": [ "Invalidation timestamp", "Successor reference where superseded", "Fields that must remain readable after termination" ] } ], "data_elements": [ { "id": "sst-d1", "name": "Record status", "description": "Current lifecycle state of the claim record.", "value_kind": "code", "cardinality": "1", "required": true, "source_refs": [ "SRC-001" ] }, { "id": "sst-d2", "name": "Status effective timestamp", "description": "Instant from which the current status applies, in RFC 3339 with seconds and offset.", "value_kind": "timestamp", "cardinality": "1", "required": true, "source_refs": [ "SRC-007" ] }, { "id": "sst-d3", "name": "Invalidation timestamp", "description": "Instant at which the record ceased to be current, where a terminal status applies.", "value_kind": "timestamp", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "sst-d4", "name": "Transition record", "description": "Actor, reason, prior status and timestamp for each status change of this record.", "value_kind": "object", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-001" ] } ], "artifacts": [ { "id": "sst-a1", "name": "Claim Lifecycle State Scheme", "description": "Definition of the permitted record states, the legal transitions between them, the role authorised for each transition, the preconditions for promotion out of draft, and the fields each terminal state must set.", "media_or_form": [ "state model definition", "controlled vocabulary / concept scheme", "tabular transition matrix" ], "serial": true, "identity_strategy": "Scheme IRI under the owning package namespace with stable per-state notations; releases are serialised with a digest, and records cite the scheme version under which their last transition was made so historical transitions stay interpretable.", "source_refs": [ "SRC-001", "SRC-012" ] } ], "inline_only_rationale": null }, { "id": "versioning-and-change-control", "name": "Versioning and Change Control", "description": "The rule distinguishing a version patch from a new claim, the immutability of asserted statement text, and the version manifest that lets any prior state of the record be reconstructed.", "source_refs": [ "SRC-001", "SRC-006", "SRC-007" ], "questions": [ { "id": "vcc-q1", "text": "Which changes are version patches to this record and which create a new claim linked by supersession?", "kind": "decision", "answer_data": [ "Change classification rule", "Worked boundary cases", "Decision record for contested changes" ] }, { "id": "vcc-q2", "text": "Which fields are immutable once the record is asserted, and how are typographic corrections handled?", "kind": "constraint", "answer_data": [ "Immutable field list", "Correction procedure", "Retained prior text" ] }, { "id": "vcc-q3", "text": "How is a specific prior version of the record cited and retrieved?", "kind": "identity", "answer_data": [ "Version identifier form", "Retrieval procedure", "Digest per version" ] }, { "id": "vcc-q4", "text": "Which precondition must a patch declare so that concurrent edits cannot silently overwrite each other?", "kind": "validation", "answer_data": [ "Prior-version precondition rule", "Conflict rejection behaviour", "Retry guidance for clients" ] } ], "data_elements": [ { "id": "vcc-d1", "name": "Version identifier", "description": "Identifier of a specific version of the claim record.", "value_kind": "identifier", "cardinality": "1", "required": true, "source_refs": [ "SRC-001" ] }, { "id": "vcc-d2", "name": "Prior version reference", "description": "Reference to the version this one revises.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "vcc-d3", "name": "Change classification", "description": "Whether a change was a patch or triggered creation of a new claim.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-006" ] } ], "artifacts": [ { "id": "vcc-a1", "name": "Claim Version Manifest", "description": "Manifest listing each version of a claim record with its version identifier, digest, prior-version reference, change classification and effective timestamp, sufficient to reconstruct any prior state of the record.", "media_or_form": [ "structured manifest", "tabular register", "graph serialization" ], "serial": true, "identity_strategy": "Manifest keyed by the primary claim identifier, with monotonically increasing version identifiers that contain no date component; each entry is append-only and carries a digest over the canonical serialization of that version.", "source_refs": [ "SRC-001", "SRC-006" ] } ], "inline_only_rationale": null }, { "id": "retraction-correction-and-status-notices", "name": "Retraction, Correction and Status Notices", "description": "References to notices issued outside this model — retractions, removals, corrections and expressions of concern — and the local status effect they trigger, with issuance and propagation remaining with the issuing publisher or authority.", "source_refs": [ "SRC-012", "SRC-003", "SRC-001" ], "questions": [ { "id": "rcn-q1", "text": "Which externally issued notice affects this claim, of what type, and who issued it?", "kind": "event", "answer_data": [ "Notice identifier and locator", "Notice type (retraction, removal, correction, expression of concern)", "Issuing body and issuance timestamp" ] }, { "id": "rcn-q2", "text": "Which local status change does the notice trigger, and is the change automatic or reviewed?", "kind": "state", "answer_data": [ "Triggered status", "Automatic or reviewed determination", "Reviewer and decision timestamp" ] }, { "id": "rcn-q3", "text": "How must a claim affected by a notice be displayed to downstream consumers?", "kind": "interoperability", "answer_data": [ "Mandatory status banner requirement", "Projection fields carrying notice type", "Prohibition on silent suppression" ] }, { "id": "rcn-q4", "text": "Who is responsible for issuing, publishing and propagating the notice itself, as opposed to referencing it here?", "kind": "authority", "answer_data": [ "Issuing and propagating parties", "This model's non-ownership statement", "Escalation route where a notice is missing or wrong" ] } ], "data_elements": [ { "id": "rcn-d1", "name": "Status notice reference", "description": "Locator of an externally issued retraction, removal, correction or expression-of-concern notice affecting the claim.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-012" ] }, { "id": "rcn-d2", "name": "Notice type", "description": "Type of the referenced notice as classified by its issuer.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-012" ] }, { "id": "rcn-d3", "name": "Notice issuance timestamp", "description": "When the issuer published the notice, in RFC 3339 with seconds and offset.", "value_kind": "timestamp", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-012" ] } ], "artifacts": [], "inline_only_rationale": "Retraction and correction notices are artifacts created, published and propagated by publishers, aggregators and issuing authorities under their own recommended practice. This model records a reference and the local status effect only; producing a notice artifact here would claim issuance authority and would duplicate a document whose canonical version lives with its issuer." } ] }, { "id": "validation-and-quality", "name": "Validation and Quality", "description": "Structural well-formedness of the claim record and the quality signals that indicate whether it is specific and testable enough to be useful.", "source_refs": [ "SRC-008", "SRC-010", "SRC-006" ], "findings": [ { "id": "well-formedness-quality-and-testability", "name": "Well-Formedness, Quality and Testability", "description": "The shapes and required-field profiles the record must satisfy, the severity grading of failures, and the quality signals — specificity, ambiguity, unresolved scope, testability — that gate promotion. Shapes are owned here; execution is not.", "source_refs": [ "SRC-008", "SRC-010", "SRC-006" ], "questions": [ { "id": "wqt-q1", "text": "Which shapes and required-field profile apply to this record given its claim type?", "kind": "validation", "answer_data": [ "Shapes package serial and digest", "Applicable node and property shapes", "Type-specific required fields" ] }, { "id": "wqt-q2", "text": "Which constraint failures are violations that block promotion and which are warnings that only annotate?", "kind": "requirement", "answer_data": [ "Severity grading per constraint", "Promotion-blocking rule", "Warning disposition procedure" ] }, { "id": "wqt-q3", "text": "Which quality signals indicate that the statement is too vague, ambiguous or under-scoped to be testable?", "kind": "quality", "answer_data": [ "Specificity and ambiguity signal values", "Unresolved-scope flags", "Testability verdict with rationale" ] }, { "id": "wqt-q4", "text": "Which component executes validation, and what does this model own as against that executor?", "kind": "process", "answer_data": [ "Processor identification and conformance level", "Ownership split statement", "Handling of processor-reported failures as distinct from violations" ] } ], "data_elements": [ { "id": "wqt-d1", "name": "Applicable shapes reference", "description": "Serial and digest of the shapes package under which the record was last validated.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-008" ] }, { "id": "wqt-d2", "name": "Conformance result", "description": "Whether the record conformed, with severity-graded results attached.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-008" ] }, { "id": "wqt-d3", "name": "Quality signal set", "description": "Specificity, ambiguity, unresolved-scope and testability signals computed for the record.", "value_kind": "collection", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-010" ] }, { "id": "wqt-d4", "name": "Promotion blocker", "description": "An outstanding violation preventing the record from leaving draft.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-008" ] } ], "artifacts": [ { "id": "wqt-a1", "name": "Claim Shapes Package", "description": "Published set of node and property shapes with severity assignments, cardinality and datatype constraints, and the per-claim-type required-field profiles, intended for execution by any conformant external validation processor.", "media_or_form": [ "shapes graph", "machine-readable constraint definition", "profile documentation" ], "serial": true, "identity_strategy": "Package IRI under the owning package namespace with a semantic-version serial and a content digest; records cite serial and digest together so that a later package release never retroactively invalidates a record that conformed at the time.", "source_refs": [ "SRC-008", "SRC-010" ] } ], "inline_only_rationale": null } ] }, { "id": "access-sensitivity-and-retention", "name": "Access, Sensitivity and Retention", "description": "Classification of claim records by sensitivity, expression of access and use conditions, and the retention, tombstoning and disposition rules for the records this model owns.", "source_refs": [ "SRC-013", "SRC-001", "SRC-012" ], "findings": [ { "id": "sensitivity-and-access-classification", "name": "Sensitivity and Access Classification", "description": "Classification of a claim record by the harm its statement could cause — unverified allegations, personal data, embargoed findings — and the expression of the resulting access and use conditions as a policy, with enforcement performed elsewhere.", "source_refs": [ "SRC-013", "SRC-002" ], "questions": [ { "id": "sac-q1", "text": "Which sensitivity class applies to this record, and which feature of the statement drives it?", "kind": "privacy", "answer_data": [ "Sensitivity class code", "Driving feature (named person, allegation, embargo, special-category data)", "Classifying agent and timestamp" ] }, { "id": "sac-q2", "text": "Which permissions, prohibitions and duties are expressed over this record, for which parties?", "kind": "access", "answer_data": [ "Permission set with actions and assignees", "Prohibition set", "Duties attached to permissions" ] }, { "id": "sac-q3", "text": "Under what constraint does an embargo lift, and what changes when it does?", "kind": "temporal", "answer_data": [ "Embargo expiry instant in RFC 3339 with offset", "Post-embargo access class", "Party responsible for confirming the lift" ] }, { "id": "sac-q4", "text": "Which component evaluates and enforces these expressed conditions, and what does this model explicitly not do?", "kind": "security", "answer_data": [ "Enforcement platform reference", "Non-ownership statement for evaluation and decision logs", "Escalation route for enforcement failures" ] } ], "data_elements": [ { "id": "sac-d1", "name": "Sensitivity class", "description": "Classification of the record by the harm potential of its statement.", "value_kind": "code", "cardinality": "1", "required": true, "source_refs": [ "SRC-013" ] }, { "id": "sac-d2", "name": "Access policy reference", "description": "Identifier of the policy expression stating permissions, prohibitions and duties over the record.", "value_kind": "reference", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-013" ] }, { "id": "sac-d3", "name": "Embargo expiry", "description": "Instant at which an embargo on the record lifts, in RFC 3339 with seconds and offset.", "value_kind": "timestamp", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-013" ] } ], "artifacts": [ { "id": "sac-a1", "name": "Claim Access Policy Expression", "description": "Machine-readable policy stating the permissions, prohibitions, duties, parties, actions and constraints that apply to a claim record or a class of claim records, expressed declaratively and carrying no enforcement semantics.", "media_or_form": [ "policy expression document", "machine-readable rule set", "graph serialization" ], "serial": true, "identity_strategy": "Policy IRI under the owning package namespace with a serial per revision and a digest; records cite the policy identifier and serial in force, and superseded policies are retained so that historical access decisions remain explicable.", "source_refs": [ "SRC-013" ] } ], "inline_only_rationale": null }, { "id": "retention-and-disposition", "name": "Retention and Disposition", "description": "How long claim records are kept, what a tombstone must carry when a record is terminated or redacted, and which party owns execution of erasure across stores and downstream copies.", "source_refs": [ "SRC-001", "SRC-012", "SRC-013" ], "questions": [ { "id": "rdd-q1", "text": "How long is a claim record retained after it reaches a terminal status, and on what trigger?", "kind": "retention", "answer_data": [ "Retention period", "Disposition trigger", "Schedule serial in force" ] }, { "id": "rdd-q2", "text": "Which fields must a tombstone carry so that downstream copies remain reconcilable?", "kind": "requirement", "answer_data": [ "Mandatory tombstone fields", "Successor reference requirement", "Redaction marker for removed statement text" ] }, { "id": "rdd-q3", "text": "Under what narrow conditions may statement text be redacted rather than retained?", "kind": "exception", "answer_data": [ "Redaction trigger conditions", "Approving role", "Record of what was redacted and why" ] }, { "id": "rdd-q4", "text": "Which party executes erasure across stores, indexes, caches and downstream copies?", "kind": "authority", "answer_data": [ "Executing party by store class", "This model's rule-only ownership statement", "Confirmation-of-execution expectation" ] } ], "data_elements": [ { "id": "rdd-d1", "name": "Retention period", "description": "Period for which the record is retained after its disposition trigger fires.", "value_kind": "duration", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-001" ] }, { "id": "rdd-d2", "name": "Disposition trigger", "description": "The condition that starts the retention clock for this record.", "value_kind": "code", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-012" ] }, { "id": "rdd-d3", "name": "Tombstone record", "description": "Surviving stub carrying identifier, terminal status, terminal timestamp, redaction marker and successor reference.", "value_kind": "object", "cardinality": "0..1", "required": false, "source_refs": [ "SRC-001" ] } ], "artifacts": [ { "id": "rdd-a1", "name": "Claim Disposition Schedule and Tombstone Specification", "description": "Schedule binding retention periods and disposition triggers to sensitivity classes and terminal statuses, together with the mandatory tombstone field set and the redaction procedure for statement text.", "media_or_form": [ "retention schedule", "structured specification", "tabular register" ], "serial": true, "identity_strategy": "Schedule IRI under the owning package namespace, serialised per release with a digest; each disposed record cites the schedule serial applied, and superseded schedules are retained for the longest retention period they governed.", "source_refs": [ "SRC-001", "SRC-012", "SRC-013" ] } ], "inline_only_rationale": null } ] } ] }, { "id": "interoperability-and-agent-operation", "name": "Interoperability and Agent Operation", "description": "How claim records project onto external vocabularies without claiming conformance, and how an agent bootstraps into the model and exchanges records safely.", "rationale": "Multiple standards model overlapping but non-identical notions of a claim: schema.org as a creative work, VC 2.0 as a subject-property-value assertion inside a credential, AIF as an I-node, nanopublications as an assertion graph, and RDF 1.2 as a triple term that may be named without being asserted. Recording these as alignments with declared losses, and providing an explicit bootstrap contract, is what makes the model operable by an agent.", "source_refs": [ "SRC-002", "SRC-003", "SRC-004", "SRC-005", "SRC-006", "SRC-009", "SRC-014" ], "layers": [ { "id": "alignment-projection-and-bootstrap", "name": "Alignment, Projection and Bootstrap", "description": "Vocabulary crosswalks with lossiness declarations and statement-level projection choices, plus the agent bootstrap contract and exchange profile.", "source_refs": [ "SRC-002", "SRC-004", "SRC-005", "SRC-006", "SRC-009" ], "findings": [ { "id": "vocabulary-alignments-and-projections", "name": "Vocabulary Alignments and Projections", "description": "Declared mappings from claim record fields to external vocabularies, the statement-level identity choice in each projection, and the explicit record of what each projection loses and where the standards conflict.", "source_refs": [ "SRC-002", "SRC-003", "SRC-004", "SRC-005", "SRC-006", "SRC-009" ], "questions": [ { "id": "vap-q1", "text": "Which external vocabulary terms does each claim record field map to, and is the mapping exact, broader or narrower?", "kind": "interoperability", "answer_data": [ "Field-to-term mapping table", "Mapping strength per row", "Crosswalk edition serial and digest" ] }, { "id": "vap-q2", "text": "In a graph projection, is the statement asserted or only named, and which construct expresses that?", "kind": "identity", "answer_data": [ "Asserted or reified determination", "Construct used (named assertion graph or reifying triple with triple term)", "Guard against converting an unasserted reifier into an asserted graph" ] }, { "id": "vap-q3", "text": "Which fields does each target profile fail to carry, and is that profile therefore declared lossy?", "kind": "validation", "answer_data": [ "Dropped field inventory per profile", "Lossy declaration per profile", "Prohibition on using a lossy projection as record of identity" ] }, { "id": "vap-q4", "text": "Where do the aligned standards disagree about what a claim is, and how is the disagreement recorded rather than papered over?", "kind": "definition", "answer_data": [ "Conflict statements between standards", "Identity-of-record designation per conflict", "Statement that alignment is not a conformance claim" ] } ], "data_elements": [ { "id": "vap-d1", "name": "Alignment mapping", "description": "A mapping from a claim record field to an external vocabulary term, with mapping strength.", "value_kind": "object", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-002" ] }, { "id": "vap-d2", "name": "Projection profile reference", "description": "Identifier of a target exchange profile with its lossiness declaration.", "value_kind": "reference", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-006" ] }, { "id": "vap-d3", "name": "Assertion mode in projection", "description": "Whether a graph projection asserts the statement or only names it via reification.", "value_kind": "code", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-009" ] }, { "id": "vap-d4", "name": "Recorded standard conflict", "description": "A documented disagreement between aligned standards about claim identity or semantics.", "value_kind": "text", "cardinality": "0..n", "required": false, "source_refs": [ "SRC-004" ] } ], "artifacts": [ { "id": "vap-a1", "name": "Claim Alignment Crosswalk", "description": "Crosswalk edition mapping claim record fields to PROV-O, schema.org Claim and ClaimReview, VC 2.0 claim structure, AIF I-node level, nanopublication graphs and RDF 1.2 constructs, with mapping strengths, per-profile lossiness declarations and recorded inter-standard conflicts.", "media_or_form": [ "crosswalk / mapping table", "machine-readable mapping set", "profile documentation" ], "serial": true, "identity_strategy": "Crosswalk IRI under the owning package namespace, edition-serialised with a digest; each projection output cites the crosswalk edition used, and editions are retained so a past projection can be re-derived exactly.", "source_refs": [ "SRC-002", "SRC-003", "SRC-004", "SRC-005", "SRC-006", "SRC-009" ] } ], "inline_only_rationale": null }, { "id": "agent-bootstrap-and-exchange-profile", "name": "Agent Bootstrap and Exchange Profile", "description": "The bootstrap contract an agent reads before operating on claim records, the order in which it must read it, and the minimum interchange profile that any exchange of claim records must satisfy.", "source_refs": [ "SRC-014", "SRC-006", "SRC-008", "SRC-007" ], "questions": [ { "id": "abe-q1", "text": "Which entry document must an agent read first, and what fields must it contain before any operation is attempted?", "kind": "requirement", "answer_data": [ "Bootstrap filename", "Mandatory field list", "Failure behaviour when a field is missing" ] }, { "id": "abe-q2", "text": "In what order must an agent read the specification, storage, interface and process references?", "kind": "process", "answer_data": [ "Ordered read sequence", "Rationale for the ordering", "Point at which referenced-model contracts must be read" ] }, { "id": "abe-q3", "text": "Which fields constitute the minimum interchange profile that any exchange of a claim record must carry?", "kind": "interoperability", "answer_data": [ "Minimum field set", "Rejection rule for incomplete exchanges", "Unknown-field forward-compatibility rule" ] }, { "id": "abe-q4", "text": "Which referenced models must an agent consult before it dereferences an evidence, review or argument identifier?", "kind": "composition", "answer_data": [ "Referenced model list with ownership statements", "Prohibited local-copy list", "Resolution failure handling" ] } ], "data_elements": [ { "id": "abe-d1", "name": "Bootstrap field set", "description": "The fields the bootstrap document must carry for an agent to operate on the register.", "value_kind": "collection", "cardinality": "1", "required": true, "source_refs": [ "SRC-014" ] }, { "id": "abe-d2", "name": "Minimum interchange profile", "description": "The field set that every exchanged claim record must carry.", "value_kind": "collection", "cardinality": "1", "required": true, "source_refs": [ "SRC-006" ] }, { "id": "abe-d3", "name": "Referenced model contract", "description": "A referenced model with a statement of what it owns and what may not be copied locally.", "value_kind": "reference", "cardinality": "1..n", "required": true, "source_refs": [ "SRC-008" ] } ], "artifacts": [ { "id": "abe-a1", "name": "AGENTS.md", "description": "Mandatory bootstrap document at the root of the model package, carrying Name, Type, Specification URL, Storage type URL, Interface URL and Processes URL together with the owning package, identity priority, scheme and shapes versions in force, and the referenced-model ownership statements.", "media_or_form": [ "bootstrap contract document", "structured front-matter record" ], "serial": true, "identity_strategy": "Fixed filename AGENTS.md at the package root, addressed by the package IRI; each release carries a serial and digest, and the file is the single entry point regardless of whether the storage projection is a document store, a graph store or a service interface.", "source_refs": [ "SRC-014", "SRC-006", "SRC-008" ] } ], "inline_only_rationale": null } ] } ] } ] }, "functions": [ { "id": "register-claim", "name": "Register claim", "description": "Create a claim record in draft from an observed or reported assertion, capturing the canonical statement, language, claim type, claimant and the two distinct timestamps.", "inputs": [ "Candidate statement text with language tag", "Claimant reference or claimant-unknown code", "Assertion timestamp or unknown-time code", "Capture context and source locator" ], "outputs": [ "Draft claim record with assigned identifier", "Atomicity verdict", "Truth-aptness verdict" ], "preconditions": [ "Content passes the truth-aptness gate", "Owning package and accountable steward are declared", "A claim type from the scheme in force can be assigned" ], "effects": [ "Creates a new record in draft state", "Records assertion time and ingestion time separately in RFC 3339 with seconds and offset", "Creates no evidence, review or argument records" ], "source_refs": [ "SRC-002", "SRC-004", "SRC-006", "SRC-007" ] }, { "id": "normalize-claim-statement", "name": "Normalize claim statement", "description": "Produce the canonical serialization and optional structured form of the statement so that equivalence comparison and digesting are deterministic.", "inputs": [ "Canonical statement text and language tag", "Normalization profile identifier" ], "outputs": [ "Canonical serialization", "Normalized subject-property-value or logical form", "Content digest as secondary equivalence key" ], "preconditions": [ "Language tag is present", "Normalization profile is published and versioned" ], "effects": [ "Enables deterministic comparison across records", "Does not alter the asserted meaning or the stored canonical text" ], "source_refs": [ "SRC-004", "SRC-006", "SRC-009" ] }, { "id": "assign-claim-identity", "name": "Assign claim identity", "description": "Select and bind the record's primary identifier by identity priority, attaching aliases for identifiers issued by other systems.", "inputs": [ "Candidate identifiers from master systems", "Namespace authority configuration", "Alias candidates discovered later" ], "outputs": [ "Primary claim identifier", "Alias set with issuing authorities" ], "preconditions": [ "Identity priority is declared by the owning package", "Namespace resolution commitment exists for the retention period" ], "effects": [ "Binds a stable identifier containing no date component", "Replaces a locally minted identifier with an alias rather than a new record when a master identifier is later discovered" ], "source_refs": [ "SRC-009", "SRC-014", "SRC-006" ] }, { "id": "qualify-claim-scope", "name": "Qualify claim scope", "description": "Attach subject, population, spatial, temporal, condition and measurand qualifiers that bound the assertion, and flag scope that remains unresolved.", "inputs": [ "Subject or population definition", "Spatial extent and validity interval", "Scenario, assumptions and known defeaters", "Measurand, value and unit where quantitative" ], "outputs": [ "Populated scope qualifier set", "Unresolved-scope flags" ], "preconditions": [ "Claim type assigned so the required-field profile is known" ], "effects": [ "Narrows the record's applicability to the stated conditions", "Raises promotion blockers where type-required qualifiers are missing" ], "source_refs": [ "SRC-010", "SRC-011" ] }, { "id": "state-calibrated-confidence", "name": "State calibrated confidence", "description": "Record the calibrated confidence or likelihood term the asserter attached, bound to a registered scale version, with a pointer to the traceable account that justifies it.", "inputs": [ "Scale identifier and version", "Confidence or likelihood term", "Traceable-account reference" ], "outputs": [ "Calibrated qualification on the record", "Scale binding reference" ], "preconditions": [ "The scale is registered and its version is in force", "The traceable-account reference resolves" ], "effects": [ "Records a stated qualification only", "Stores no evidence-body appraisal grade, which remains dereferenced from the evidence model" ], "source_refs": [ "SRC-010" ] }, { "id": "bind-evidence-reference", "name": "Bind evidence reference", "description": "Attach a qualified reference from the claim to an evidence item, carrying the binding role, directionality and locally stated weight and nothing else.", "inputs": [ "Evidence item identifier in the referenced evidence model", "Binding role code", "Directionality and locally stated binding weight" ], "outputs": [ "Evidence binding edge", "Binding status" ], "preconditions": [ "Target identifier resolves", "Role code exists in the registered code list" ], "effects": [ "Adds or updates a local edge only", "Never writes to the evidence model and never copies evidence description, provenance or appraisal into the claim record" ], "source_refs": [ "SRC-005", "SRC-006", "SRC-014" ] }, { "id": "assert-claim-relation", "name": "Assert claim relation", "description": "Record a typed relation between two claim records — entailment, specialization, presupposition, conflict, preference or supersession — with its ground and asserter.", "inputs": [ "Source and target claim identifiers", "Relation code and ground", "Asserting agent and timestamp" ], "outputs": [ "Typed relation edge", "Registration event record" ], "preconditions": [ "Both claim records exist", "Relation code is registered", "Supersession would not create a cycle" ], "effects": [ "Records the relation as a qualified edge", "Registers conflict without resolving it and preference without applying it in any evaluation" ], "source_refs": [ "SRC-005", "SRC-001", "SRC-014" ] }, { "id": "transition-claim-status", "name": "Transition claim status", "description": "Move a claim record between permitted lifecycle states, setting effective and terminal timestamps and any successor reference.", "inputs": [ "Claim identifier and target status", "Acting role, reason and effective instant" ], "outputs": [ "Updated record status", "Transition record" ], "preconditions": [ "The transition is permitted by the state scheme in force", "The acting role is authorised for that transition", "Promotion preconditions are satisfied when leaving draft" ], "effects": [ "Changes state and writes the transition record for this model's own record", "Sets an invalidation timestamp and successor reference on terminal transitions" ], "source_refs": [ "SRC-001", "SRC-007", "SRC-012" ] }, { "id": "record-status-notice-reference", "name": "Record status notice reference", "description": "Attach a reference to an externally issued retraction, removal, correction or expression-of-concern notice and apply the declared local status effect.", "inputs": [ "Notice locator, type and issuer", "Notice issuance timestamp", "Review determination where the effect is not automatic" ], "outputs": [ "Notice reference on the claim record", "Resulting local status" ], "preconditions": [ "The notice was issued by an authority outside this model", "The notice type maps to a declared local status effect" ], "effects": [ "Records the reference and applies the local status effect", "Does not issue, publish or propagate the notice, which remains the issuing publisher's or authority's responsibility" ], "source_refs": [ "SRC-012", "SRC-003" ] }, { "id": "validate-claim-record", "name": "Validate claim record", "description": "Check a claim record against the published shapes package and the claim-type required-field profile, producing a severity-graded conformance report.", "inputs": [ "Claim record", "Shapes package serial and digest", "Claim-type profile version" ], "outputs": [ "Conformance report with severity-graded results", "Promotion blocker list" ], "preconditions": [ "Shapes package is published with a digest", "The record's claim type is assigned" ], "effects": [ "Blocks promotion out of draft while violations remain", "Owns shapes and profiles only; execution is performed by an external conformant validation processor whose engine and run logs are not owned here" ], "source_refs": [ "SRC-008", "SRC-010" ] }, { "id": "classify-and-express-access", "name": "Classify sensitivity and express access conditions", "description": "Assign a sensitivity class to a claim record and express the resulting permissions, prohibitions and duties as a declarative policy.", "inputs": [ "Claim record and statement text", "Sensitivity criteria", "Policy template and party references" ], "outputs": [ "Sensitivity class", "Access policy expression reference" ], "preconditions": [ "Statement text has been reviewed for personal data and allegation content", "Classifying role is authorised" ], "effects": [ "Records a class and a policy expression", "Performs no evaluation, decisioning or enforcement, which belong to the adopting Dimension's platform" ], "source_refs": [ "SRC-013", "SRC-002" ] }, { "id": "project-claim-record", "name": "Project claim record", "description": "Emit a claim record in a target exchange profile using the crosswalk edition in force, declaring what the projection drops.", "inputs": [ "Claim record", "Target profile identifier", "Crosswalk edition serial and digest" ], "outputs": [ "Profile-specific projection", "Lossiness declaration listing dropped fields" ], "preconditions": [ "A crosswalk edition covering the target profile exists", "The assertion mode for graph projections has been determined" ], "effects": [ "Produces an exchange artifact marked with the crosswalk edition used", "Marks a lossy projection as unusable as a record of identity and never converts an unasserted reifier into an asserted graph" ], "source_refs": [ "SRC-002", "SRC-004", "SRC-005", "SRC-006", "SRC-009" ] }, { "id": "dispose-claim-record", "name": "Dispose claim record", "description": "Apply the disposition schedule to a terminated record, writing a tombstone, redacting statement text where an erasure obligation requires it, and notifying dependents.", "inputs": [ "Claim identifier", "Disposition trigger and schedule serial", "Erasure obligation determination where applicable" ], "outputs": [ "Tombstone record", "Dangling-reference notices to dependent claims" ], "preconditions": [ "A terminal status has been reached or an erasure obligation has been established", "Approving role has authorised any redaction" ], "effects": [ "Retains identifier, terminal status, terminal timestamp and successor reference in the tombstone", "Never deletes referenced evidence, reviews or argument structures; removes only the local reference edge", "Defines the rule only: execution across stores, indexes, caches and downstream copies is owned by the adopting Dimension" ], "source_refs": [ "SRC-001", "SRC-012", "SRC-013" ] } ], "composition": [ { "target": "WM-KNW-001", "relation": "CHILD", "purpose": "Inherit generic knowledge-asset identity, register governance, stewardship and licensing from the parent so this model specializes only what makes a record truth-apt. The separation of an assertion from the record metadata about it, as in PROV and the nanopublication pattern, is what makes this inheritance clean rather than duplicative.", "required": true, "source_refs": [ "SRC-001", "SRC-006" ] }, { "target": "WM-KNW-008", "relation": "REFERENCE", "purpose": "Carry the evidence item identifier, binding role, directionality and a locally stated binding weight so a claim can be shown as supported or challenged. Evidence description, provenance, appraisal grades and evidence lifecycle remain owned by the target and are dereferenced, never copied.", "required": false, "source_refs": [ "SRC-005", "SRC-006", "SRC-014" ] }, { "target": "W3C PROV-O (prov:Entity, wasAttributedTo, wasDerivedFrom, hadPrimarySource, wasRevisionOf, invalidatedAtTime, prov:Bundle)", "relation": "ALIGN", "purpose": "Express record attribution, derivation, primary source, revision and invalidation using established provenance terms. Generic Activity and Agent modelling and provenance query semantics remain with PROV; this is alignment, not a conformance claim.", "required": false, "source_refs": [ "SRC-001" ] }, { "target": "schema.org Claim and ClaimReview", "relation": "ALIGN", "purpose": "Map canonical statement text, appearance and firstAppearance, and express reviews of a claim as external ClaimReview references. schema.org defines no formal claim-to-claim relations, so the model's typed relation layer is additional and must be declared as such in the crosswalk.", "required": false, "source_refs": [ "SRC-002", "SRC-003" ] }, { "target": "W3C Verifiable Credentials Data Model 2.0 claim (subject-property-value)", "relation": "ALIGN", "purpose": "Align the normalized statement form with the VC subject-property-value structure so claims can be carried in credentials. Issuer proof, credentialStatus and validity windows belong to credential transport and must never be read as claim truth or claim lifecycle.", "required": false, "source_refs": [ "SRC-004" ] }, { "target": "Argument Interchange Format (AIF) upper ontology", "relation": "ALIGN", "purpose": "Map a claim record to the AIF I-node level so claims can participate in argument graphs. Applications of rules of inference, conflict and preference (RA, CA, PA) and schematic form nodes stay with the argumentation model.", "required": false, "source_refs": [ "SRC-005" ] }, { "target": "Nanopublication head / assertion / provenance / publicationInfo graph pattern", "relation": "ALIGN", "purpose": "Adopt the separation of the minimal assertion from its provenance and publication metadata as the reference projection for graph exchange, and adopt the guidance to keep assertions minimal as the basis of the atomicity rule.", "required": false, "source_refs": [ "SRC-006" ] }, { "target": "W3C RDF 1.2 triple terms, reifying triples and named graphs", "relation": "ALIGN", "purpose": "Provide statement-level identity in graph projections and preserve the distinction between naming a statement and asserting it, so that an unasserted reifier is never silently promoted to an asserted assertion graph.", "required": false, "source_refs": [ "SRC-009" ] }, { "target": "W3C SHACL shapes and validation reports", "relation": "REFERENCE", "purpose": "Publish the model's shapes and required-field profiles for execution by any conformant processor. Validation execution, processor failures and run logs are owned by the processor and its operator, not by this model.", "required": false, "source_refs": [ "SRC-008" ] }, { "target": "W3C ODRL Information Model 2.2", "relation": "REFERENCE", "purpose": "Express permissions, prohibitions, duties and constraints over claim records declaratively. ODRL defines no enforcement or evaluation mechanism, so no decisioning, enforcement or policy-decision logging is claimed here.", "required": false, "source_refs": [ "SRC-013" ] }, { "target": "IPCC calibrated uncertainty language (evidence and agreement, confidence and likelihood scales)", "relation": "ALIGN", "purpose": "Bind the stated confidence and likelihood terms on a claim to a published calibrated scale with a traceable account, and adopt the requirement to specify variables, region, period and conditions as the basis of the scope-qualifier layer.", "required": false, "source_refs": [ "SRC-010" ] }, { "target": "JCGM 200:2012 VIM (measurand, measurement uncertainty, coverage interval, metrological traceability)", "relation": "ALIGN", "purpose": "Use agreed metrology terms for the quantitative content of measurement claims. Measurement procedure, calibration and the metrological traceability chain remain with a metrology or observation model.", "required": false, "source_refs": [ "SRC-011" ] }, { "target": "NISO RP-45-2024 CREC retraction, removal and expression-of-concern metadata", "relation": "ALIGN", "purpose": "Align the shape of externally issued status-notice references and the requirement that an affected item's status be visible to readers. Issuance, publication and propagation of notices stay with publishers, aggregators and hosts.", "required": false, "source_refs": [ "SRC-012" ] }, { "target": "IETF RFC 3339 date-time", "relation": "ALIGN", "purpose": "Constrain every time value in the model to a date-time with seconds and an explicit offset or Z, and preserve the RFC's distinction between an unknown local offset and UTC.", "required": true, "source_refs": [ "SRC-007" ] }, { "target": "FAIR Guiding Principles F1, I3 and R1.2", "relation": "ALIGN", "purpose": "Require globally unique persistent identifiers for claim records, qualified rather than bare references for claim relations and evidence bindings, and detailed record provenance.", "required": false, "source_refs": [ "SRC-014" ] }, { "target": "Argumentation / inference-scheme model (not yet registered - declared gap)", "relation": "REFERENCE", "purpose": "Placeholder for the sibling that must own argument-scheme definition, inference, conflict and preference rule application and any evaluation of argument acceptability. Until it is registered this remains a declared gap; no local ownership of those semantics is asserted and only a role reference is held.", "required": false, "source_refs": [ "SRC-005" ] } ], "serviceLayers": { "dimension": { "owner_package_requirements": [ "The adopting Dimension MUST name one owning package for the claim register and one accountable claim steward before any record is promoted out of draft; an unowned register may hold drafts only.", "The owning package MUST declare which authority's identifiers are treated as master for its claim population, and MUST NOT mint local UUIDs or ULIDs where a master identifier already exists.", "The owning package MUST publish and version its claim-type scheme, its calibrated uncertainty scale binding, its lifecycle state scheme and its shapes package; records created under a superseded version remain valid and are migrated by explicit patch, never by silent reinterpretation.", "The owning package MUST NOT store evidence descriptions, evidence appraisal grades, review ratings or argument-scheme definitions locally; it stores resolvable references and the binding parameters this model owns.", "The owning package MUST declare the enforcement platform that evaluates its access policies and the party that executes disposition, since this model owns neither." ], "namespace_guidance": "Claim identifiers live under a namespace the owning package controls and can keep resolvable for the record's full retention period. An identifier MUST NOT be reused for a different proposition, MUST NOT encode status, confidence, claim type or any date component, and MUST survive restatement and translation of the claim text. Alignment namespaces for provenance, schema, argumentation, nanopublication, shapes and policy vocabularies are imported as external terms and are never re-minted or shadowed under the local namespace; where a local term is needed because no external term exists, it is minted locally and declared in the crosswalk as an extension rather than presented as a standard term.", "registry_links": [ "vr.wm-knw-007 is the registry entry of record for this model, at nav path NAV.INF.KNW.CLM under domain tags INF.KNW.CLM.", "WM-KNW-001 is the registered parent and supplies generic knowledge-asset governance through the CHILD link.", "WM-KNW-008 is the registered REFERENCE target for evidence and citations; the relation ledger at planning/VERCY-MODEL-RELATIONS.csv is authoritative and no local structure may contradict it.", "The argumentation / inference-scheme owner is unregistered; the gap MUST be raised against the registry before any inference-scheme content is added anywhere in this model." ] }, "canon_and_patch": { "canonicalization_rules": [ "Canonicalize statement text by trimming leading and trailing whitespace, collapsing internal runs of whitespace and normalizing Unicode to NFC, while preserving case, punctuation, diacritics and the language tag; do not lowercase, stem or strip stop words, because negation, named entities and hedges are meaning-bearing.", "Canonicalize order-insensitive collections (claim relations, evidence bindings, alignments, appearance locators) by sorting on target identifier then relation or role code, so serialization order can never change a record digest.", "Canonicalize all time values to RFC 3339 with seconds and an explicit offset; retain the original offset rather than converting to UTC, because the offset carries assertion-context meaning, and preserve the distinction between an unknown local offset and UTC.", "Treat unknown, empty and not-applicable as distinct: omit the field when a value is unknown, use the scheme's explicit not-applicable code where one exists, and never write a placeholder string such as 'n/a' or 'unknown' into a value field.", "Canonicalize references as identifier plus relation qualifier; a bare identifier without its qualifier is not canonical and fails validation." ], "patch_rules": [ "The canonical statement text of a record in asserted state is immutable: a change of meaning creates a new claim record linked by supersedes, while a typographic correction is a version patch that keeps the identifier and retains the prior text in the version manifest.", "Patches are expressed as field-level operations against the canonical form with a precondition naming the prior version identifier; blind whole-document replacement and patches without a version precondition are rejected.", "Adding, changing or removing an evidence binding, an assessment reference, a claim relation or a calibrated confidence value patches this record only and never writes into the evidence model, the reviewing system or the argumentation owner.", "A patch that would alter scope qualifiers in a way that changes which subjects, places, periods or conditions the claim covers is a meaning change and follows the new-record rule, not the patch rule.", "Every patch increments the version identifier and records actor, reason and RFC 3339 timestamp with offset." ], "compatibility_rules": [ "Adding an optional field, a new relation code, a new claim-type concept or a new alignment is backward compatible; narrowing cardinality, adding a required field, or retiring a claim-type, status or confidence-scale term is breaking and requires a new major profile version.", "Consumers MUST ignore unknown fields and MUST NOT infer that a property is absent from the record merely because a projection profile does not carry it.", "A projection that cannot carry the primary identifier, the canonical statement, the record status and the scope qualifiers MUST be declared lossy in the crosswalk and MUST NOT be used as a system of record.", "Deprecated scheme terms remain resolvable with a successor pointer for at least the longest retention period of any record that used them." ] }, "artifact_rules": { "identity_priority": [ "Authoritative master-system identifier issued by the system of record for the claim population, such as a standards-body clause or provision identifier, a regulator's decision reference, a registry claim number or a publisher-assigned claim key.", "Governed global identifier or IRI, such as a persistent HTTP IRI under a controlled namespace, a DOI, or a content-derived global name issued by a governed publication network.", "UUID or ULID minted by the adopting Dimension, used only when no master or governed global identifier exists, and replaced by an alias rather than a new record if one is later discovered.", "Content digest over the canonical normalized statement, permitted only as a secondary equivalence key and never as the record's primary identifier, because one proposition may have several texts and one text may express different claims in different contexts." ], "timestamp_rule": "Every time value in this model is an RFC 3339 date-time that includes seconds and an explicit numeric offset or Z; an unknown local offset is written as -00:00 and is not equivalent to Z. Event time and observation time are recorded separately and are never conflated: the assertion time (when the claimant asserted the proposition) and the claim's validity interval (the period for which the content is asserted true) are distinct from the ingestion or observation time (when this register captured the claim) and from the status effective time (when a lifecycle transition took effect). A date value alone is never an identifier and never a substitute for a timestamp where an instant is required.", "serial_naming_rule": "Serial artifacts - claim-type scheme, uncertainty scale registry, lifecycle state scheme, shapes package, alignment crosswalk, stewardship declaration, disposition schedule and AGENTS.md - are named as owning-package, artifact slug and a monotonically increasing major.minor.patch sequence. The sequence never encodes a date or a status, is never reset, and a superseded serial is retained and marked superseded with a successor pointer rather than renamed, reused or deleted.", "integrity_rule": "Every artifact carries a named-algorithm content digest computed over its canonical serialization together with the identifier of the record set it governs. A claim record cites both the artifact serial and its digest; a digest mismatch invalidates the citation and blocks promotion rather than silently resolving to the current artifact version. Provenance bundles and version manifests are append-only, so integrity failures surface as an added correcting entry, never as an in-place rewrite." }, "policies": [ "Assert-what-is-asserted: the register records that a proposition was asserted, by whom, under what qualification and with what standing. It never records this model's own view of the proposition's truth, and no field - not confidence, not authority standing, not a review reference, not a credential status - may be read or projected as a truth verdict.", "Reference-not-replicate: evidence, citations, reviews, argument schemes, mandates, retraction notices and enforcement decisions are referenced by resolvable identifier. Only the binding parameters this model owns may be stored locally, and any field that begins to describe a referenced object is a boundary violation to be migrated to the owning model.", "One record, one proposition: compound or composite statements are split before promotion, and a record failing the atomicity test cannot leave draft.", "No silent reinterpretation: changing a scheme, scale, state model or shapes version never retroactively changes the meaning or conformance of existing records; migration is explicit, versioned and recorded.", "Separation of expression from execution: this model expresses shapes, policies, disposition rules and status effects; validation processors, enforcement platforms, publishers and storage operators execute them and own their run logs." ], "crud": { "read": [ "Reading a claim record returns the primary identifier, canonical statement, language, claim type, scope qualifiers, stated confidence and status; evidence, review and argument information is returned as resolvable references, never as embedded copies.", "Reads MUST support filtering and must always expose status, so that consumers cannot surface retracted, withdrawn or superseded claims as current; a projection that omits status is non-conforming and may not be used as a system of record.", "Reading a superseded record returns its successor reference and reading a superseding record returns what it supersedes, so revision chains are traversable in both directions.", "Reading a record whose statement text is withheld under sensitivity rules still returns identifier, claim type, status and the fact of withholding, so downstream copies remain reconcilable." ], "create": [ "A claim record may be created only with a canonical statement, a language tag, a truth-aptness verdict, a claim type, a claimant reference or an explicit claimant-unknown code, an assertion timestamp or explicit unknown-time code, and an owning package with an accountable steward.", "Identity is assigned at creation in identity-priority order; a locally minted identifier discovered later to have a master equivalent is superseded by an alias binding, not by creating a second record.", "Creation records ingestion or observation time separately from assertion time, both as RFC 3339 values with seconds and an offset.", "Creation never creates records in referenced models; an evidence item or review that does not yet exist is raised with its owner rather than stubbed locally." ], "update": [ "Scope qualifiers, calibrated confidence, relations, bindings, alignments and sensitivity classification are updatable in place under the patch rules, each update incrementing the version identifier and recording actor, reason and timestamp.", "Status changes are performed only through the status transition function and are constrained by the lifecycle state scheme in force; a direct write to the status field is rejected.", "The canonical statement of an asserted record is not updatable in meaning; meaning changes create a new record linked by supersedes.", "Updates never mutate a referenced model's records; a needed change to evidence, a review or an argument structure is raised as a request to that model's owner and tracked as an open referral." ], "delete": [ "Default is non-destructive: an asserted claim record is never physically removed. It is transitioned to a terminal status, keeps its primary identifier, and retains a tombstone carrying the identifier, terminal status, terminal timestamp, redaction marker where applicable and successor reference where superseded.", "Hard deletion is permitted only for records never published outside the owning package, or where an erasure obligation applies to personal data embedded in the statement; in that case the statement text is redacted while the identifier, claim type, status and tombstone survive so downstream copies remain reconcilable.", "Retention periods and disposition triggers are bound to sensitivity class and terminal status by the disposition schedule; the schedule serial applied is recorded on the disposed record, and provenance bundles and version manifests inherit the record's retention rather than being purged independently.", "This model owns the disposition rule and the tombstone shape only. Execution of erasure across stores, indexes, caches, backups and downstream copies is owned by the adopting Dimension's data-protection and storage policy; issuance and propagation of retraction or removal notices in publishing channels is owned by the issuing publisher, aggregator or authority under NISO CREC practice; deletion of referenced evidence records is owned by WM-KNW-008 and deletion of review records by the reviewing system.", "Deleting or disposing of a claim record never deletes referenced evidence, reviews or argument structures; only the local reference edge is removed, and dependent claims holding relations to the disposed record are notified so their edges are marked dangling rather than silently broken." ] }, "roles": [ { "name": "Claim Steward", "responsibilities": [ "Curate canonical statements and enforce the one-record-one-proposition rule", "Assign claim type and populate the scope qualifiers the type requires", "Promote records out of draft only when validation and quality gates pass", "Record restatement, translation and equivalence links with the test applied" ] }, { "name": "Assertion Registrar", "responsibilities": [ "Record claimant identity, capacity, commitment level and appearance locators accurately", "Record asserting authority, mandate reference and standing without treating standing as truth", "Maintain the separation between assertion time, validity interval and ingestion time", "Correct attribution errors promptly and record the correction" ] }, { "name": "Register Owner and Model Governor", "responsibilities": [ "Own the claim-type scheme, uncertainty scale registry, lifecycle state scheme, shapes package and crosswalk, including their serials and digests", "Approve breaking profile changes and publish migration instructions", "Adjudicate whether a change is a patch or a new claim when stewards disagree", "Maintain the referenced-model contracts and raise boundary violations to the registry" ] }, { "name": "Data Protection and Access Officer", "responsibilities": [ "Classify sensitivity for records naming identifiable persons, unproven allegations and embargoed findings", "Express access and use conditions as declarative policy and keep them current", "Own the disposition schedule and approve any redaction of statement text", "Delegate evaluation and enforcement to the adopting Dimension's platform and confirm execution of disposition" ] }, { "name": "Interoperability Engineer", "responsibilities": [ "Maintain crosswalk editions and declare per-profile lossiness", "Determine the assertion mode for graph projections and guard against promoting an unasserted reifier to an asserted graph", "Monitor upstream standard changes and record newly discovered inter-standard conflicts", "Maintain the minimum interchange profile and the AGENTS.md bootstrap contract" ] } ], "access": { "default_rule": "Claim records are readable to the owning package and to consumers granted the register's read scope, with status always exposed. A record whose statement names or renders identifiable a natural person and whose content is an unverified allegation defaults to restricted read until a sensitivity class has been assigned; write access to statement text, status and sensitivity class is restricted to the roles named in the stewardship declaration.", "scopes": [ "bundle", "layer", "finding", "artifact" ], "exceptions": [ "Embargoed claims, such as pre-publication findings or matters under legal restriction, are withheld from general read until the embargo expiry instant, stored in RFC 3339 with an explicit offset, and the party responsible for confirming the lift is named.", "Records containing special-category personal data or unproven allegations may have their canonical statement text withheld while identifier, claim type, status and the fact of withholding remain readable.", "Retracted, withdrawn and superseded claims remain readable with a mandatory status indication; suppressing them entirely is prohibited because downstream copies must remain reconcilable with the register.", "Break-glass read for legal, safety or regulatory review is permitted with a recorded reason, a named approver and a time limit set by the adopting Dimension.", "Artifact-scope access may be granted independently of record-scope access, since schemes, scales, shapes and crosswalks are usually publishable even where individual claim records are restricted." ], "audit_requirements": [ "Every status transition, statement patch, sensitivity reclassification, relation or binding change and disposition action on this model's own records is recorded with actor, reason, prior version and an RFC 3339 timestamp with offset in the model's version manifest and provenance bundle.", "This model does not own runtime access logs, policy-decision logs, validation processor run logs or fact-check workflow trails; those are produced, retained and disposed of by the adopting Dimension's enforcement platform, the validation processor operator and the reviewing systems referenced from this model.", "Audit entries in the version manifest and provenance bundle are append-only and inherit the retention of the record they describe; they are tombstoned rather than deleted, and a correction is added as a new entry rather than by rewriting an existing one.", "Break-glass reads and redaction approvals are recorded with approver identity and expiry, and are reviewed by the Data Protection and Access Officer on the cycle the adopting Dimension declares." ] }, "agents_bootstrap": { "filename": "AGENTS.md", "required_fields": [ "Name", "Type", "Specification URL", "Storage type URL", "Interface URL", "Processes URL", "Owning package and accountable steward", "Identity priority and controlled namespace", "Scheme, scale, state model and shapes package serials in force", "Referenced models and what each owns, including the prohibition on copying evidence, review or argument content locally", "Minimum interchange profile and declared lossy projections" ], "read_order": [ "AGENTS.md at the package root, before any read or write operation is attempted", "Specification URL, for scope, in-scope and out-of-scope lists and boundary notes, so that ownership limits are known before data is touched", "Storage type URL, for the canonical form and the projection in use, since storage is a projection and not the semantics", "Interface URL, for read and write operations, status filtering and the version-precondition requirement on patches", "Processes URL, for registration, scope qualification, status transition, validation, projection and disposition procedures", "Referenced-model contracts for WM-KNW-001 and WM-KNW-008, and the declared argumentation gap, before dereferencing any evidence, review or argument identifier" ] } }, "coverage": { "claim": "Single-provider Claude draft for WM-KNW-007 covering the claim record as a governed knowledge object: canonical statement and logical form, identity and equivalence, subject/spatial/temporal/conditional scope, quantified content and calibrated uncertainty, the assertion act and asserting authority, record provenance, ownership and rights, typed claim-to-claim relations, qualified bindings to evidence, reviews and argument structures, lifecycle, versioning and externally issued notices, validation shapes, sensitivity, access and disposition, and interoperability projections — 6 bundles, 12 layers, 28 findings, 112 questions, 12 artifacts and 13 functions grounded in 14 declared sources. Two checklist dimensions remain open gaps (argument/inference ownership; privacy legal basis), several typologies are uncited design proposals, three findings have no operational function, and no independent second-provider review exists. No universal, domain-complete or metaphysical completeness is claimed.", "confidence": "medium", "checklist": [ { "dimension": "identity", "status": "covered", "notes": "Identity priority places the authoritative master-system identifier first, then a governed global identifier or IRI, then a locally minted UUID/ULID, with content digests demoted to a secondary equivalence key. No identifier may contain a date, status or confidence component. Grounded in RDF 1.2 IRI identity, FAIR F1 and nanopublication practice (SRC-009, SRC-014, SRC-006)." }, { "dimension": "lifecycle", "status": "covered", "notes": "A published state scheme defines permitted states, transitions, authorised roles, promotion preconditions and terminal effects, with PROV invalidation semantics for termination and an explicit new-claim versus new-version rule. Externally issued retraction and correction notices are referenced, not issued here (SRC-001, SRC-012)." }, { "dimension": "relationships", "status": "covered", "notes": "Entailment, specialization, presupposition, conflict, preference and supersession are recorded as qualified typed edges, matching the registry's TYPED-EDGES default. Registration is explicitly not adjudication: conflict resolution and preference application belong to the argumentation model (SRC-005, SRC-001, SRC-014)." }, { "dimension": "temporal", "status": "covered", "notes": "Assertion time, validity interval, ingestion/observation time and status effective time are modelled as distinct values, all RFC 3339 with seconds and an explicit offset, with the unknown-local-offset case preserved as -00:00 (SRC-007, SRC-010)." }, { "dimension": "provenance", "status": "covered", "notes": "Derivation, quotation, primary source, revision, generating agent and extraction method are captured, with extraction confidence kept distinct from the claim's own stated confidence, and materialised as an append-only provenance bundle following PROV and the nanopublication provenance graph (SRC-001, SRC-006)." }, { "dimension": "ownership", "status": "covered", "notes": "Owning package, accountable steward, role-to-field write matrix, licence and rights for quoted text, and a succession rule on package retirement are declared in a serialised stewardship declaration (SRC-014, SRC-013)." }, { "dimension": "validation", "status": "covered", "notes": "A published shapes package with severity-graded constraints and per-claim-type required-field profiles gates promotion out of draft. The model owns shapes and profiles; execution is delegated to an external conformant processor whose engine and run logs are outside the boundary (SRC-008)." }, { "dimension": "access", "status": "covered", "notes": "Sensitivity classes, a restricted default for unverified claims naming identifiable persons, embargo handling, break-glass exceptions and declarative policy expression are defined. ODRL is used for expression only; it defines no enforcement mechanism and none is claimed (SRC-013, SRC-002)." }, { "dimension": "retention and deletion", "status": "covered", "notes": "Non-destructive default with mandatory tombstone fields, narrow redaction conditions for embedded personal data, a serialised disposition schedule, and an explicit statement that this model owns the rule while the adopting Dimension executes erasure, publishers execute notices and referenced models own deletion of their own records (SRC-001, SRC-012, SRC-013)." }, { "dimension": "interoperability", "status": "covered", "notes": "A serialised crosswalk maps fields to PROV-O, schema.org, VC 2.0, AIF, nanopublication and RDF 1.2 with per-profile lossiness declarations, an explicit asserted-versus-reified projection decision, and recorded inter-standard conflicts. Alignment is never presented as conformance (SRC-002, SRC-003, SRC-004, SRC-005, SRC-006, SRC-009)." }, { "dimension": "classification", "status": "covered", "notes": "A truth-aptness gate rejects questions, commands, performatives and bare preferences, and a governed claim-type scheme drives type-specific required fields. Grounded in schema.org's factually oriented Claim definition and the IPCC requirement to specify what is being assessed (SRC-002, SRC-010)." }, { "dimension": "scope and qualification", "status": "covered", "notes": "Subject and population range with inclusion/exclusion criteria and aggregation level, spatial and jurisdictional extent, validity interval, scenario, assumptions and acknowledged defeaters are all modelled, following the IPCC instruction to specify variables, region, period and conditions (SRC-010)." }, { "dimension": "measurement and uncertainty", "status": "covered", "notes": "Measurand, value, unit, stated coverage interval with coverage probability and comparator baseline are modelled, alongside a registered calibrated confidence and likelihood scale with a traceable-account pointer. Measurement procedure and traceability chains are out of scope (SRC-011, SRC-010)." }, { "dimension": "authority", "status": "covered", "notes": "Mandate reference, recognised jurisdiction and official/advisory/unofficial standing are recorded, with an explicit constraint preventing standing from being projected as a truth verdict. Mandates themselves remain instruments of their granting bodies (SRC-004, SRC-002, SRC-010)." }, { "dimension": "evidence binding boundary", "status": "covered", "notes": "Only target identifier, binding role, directionality, locally stated weight and binding status are held. Evidence description, provenance and appraisal grades are explicitly excluded, and a dedicated question tests for appraisal fields leaking into the calibrated-confidence finding (SRC-005, SRC-006, SRC-014)." }, { "dimension": "argument and inference ownership", "status": "gap", "notes": "No sibling model owning argument-scheme definition, inference/conflict/preference rule application or argument evaluation is registered. The model holds only an AIF I-node role reference and a declared placeholder composition link. Until that owner exists, any consumer wanting evaluated arguments has no in-registry target, and this node is marked a gap rather than presented as canonical (SRC-005)." }, { "dimension": "privacy and personal data", "status": "gap", "notes": "Sensitivity classification, restricted defaults and redaction procedures are defined, but the underlying legal duties (accuracy, rectification, erasure) could not be cited from a retrievable primary legal source because the EUR-Lex client requests were blocked. This dimension rests on a regional assumption rather than a cited requirement and must be confirmed by the adopting Dimension's counsel." } ], "known_omissions": [ "The IPCC uncertainty guidance note (SRC-010) was read from a mirrored copy because the canonical ipcc.ch URL returned HTTP 403 to the research client; the cited URL is the canonical location and the content was verified against the mirror.", "The JCGM 200:2012 PDF did not extract cleanly; VIM term coverage (measurand, measurement result, measurement uncertainty, coverage interval, coverage probability, metrological traceability) was confirmed from the BIPM catalogue description rather than full-text extraction, so the metrology link is declared an alignment and not a conformance claim.", "EUR-Lex requests for Regulation (EU) 2016/679 were blocked, so data-protection duties are recorded as regional assumptions rather than cited requirements; no GDPR source is cited.", "No normative registry of claim-to-claim relation types was found. The proposed relation codes synthesise AIF conflict and preference semantics with PROV revision semantics and are a design proposal, not a standard.", "The commitment-level and illocutionary-force typology in the assertion-act finding has no cited normative source and is a design proposal drawing on the schema.org distinction between a claim and its review context.", "Deontic and normative claims (obligations, permissions, recommendations) are covered only by the modality field; a full normative-statement model, including the relation between a norm and its enforcement, is not attempted here and ODRL is cited for policy about records rather than for norms as claim content.", "Thresholds for cross-language claim equivalence, and any automated similarity method for deciding sameness, are left to the adopting Dimension; only the requirement to record the test applied is specified.", "Aggregation of many claims into positions, consensus statements or knowledge-base states is not modelled; it would require the unregistered argumentation owner or a separate aggregate model.", "The AIF specification PDF returned an imprecise automated summary on first extraction; RA, CA and PA node meanings were re-verified as applications of rules of inference, conflict and preference before use." ], "conflicts": [ "schema.org models a Claim as a CreativeWork with text, name and appearance properties and defines no formal claim-to-claim relations, while AIF models the same content as an I-node inside a typed argument graph. A creative work and a proposition are different identity bearers, so the crosswalk must state which is the record of identity for a given exchange; this model treats the proposition as the identity bearer.", "W3C VC 2.0 uses 'claim' for a subject-property-value assertion inside a credential whose standing depends on issuer, proof, validity window and credentialStatus, whereas this model treats a claim as truth-apt content independent of transport. Credential expiry, revocation or status MUST NOT be projected as claim falsity or as a claim lifecycle transition.", "Nanopublication practice treats the assertion as minimal, immutable and content-addressed, whereas ClaimReview workflows expect claimReviewed to be an editable summary of prose. The model resolves this by making asserted statement meaning immutable and treating reviewer or curator summaries as separate restatement records rather than edits.", "IPCC calibrated confidence is a property of an assessed finding derived from an appraised evidence body, while evidence-body certainty ratings sit on the evidence itself. Placing a confidence term on a claim without a resolvable traceable account risks double-counting appraisal that WM-KNW-008 owns; the model therefore stores the term and the pointer but forbids storing appraisal grades.", "RDF 1.2 reifying triples allow a statement to be named and described without being asserted, whereas a nanopublication assertion graph is asserted by virtue of publication. A projection that converts a reifier into an assertion graph silently changes the record's assertive force and is prohibited.", "NISO CREC assumes a scholarly publishing chain of publishers, aggregators and hosts, while claims in this register may originate from speech, regulation or data pipelines where no such chain exists. Where no issuer of notices exists, the model has no external source of retraction and must rely on its own status transitions, which is a weaker guarantee and must be declared." ], "regional_assumptions": [ "EU/EEA: accuracy, rectification and erasure duties over personal data embedded in claim statements are assumed to apply, and the redaction and tombstone rules are designed to accommodate them; the legal text could not be retrieved by the research client and this is therefore an assumption, not a cited requirement.", "Defamation and unproven-allegation exposure varies materially by jurisdiction. The restricted-read default for unverified claims naming identifiable persons is a conservative design choice, not a cited legal requirement, and the adopting Dimension must confirm it against local law.", "NISO CREC is a US-originating recommended practice with international participation and is oriented to scholarly publishing; adoption by non-scholarly publishers, regulators or media organisations is not assumed.", "IPCC calibrated language is authoritative within climate assessment. Reusing its confidence and likelihood scales in other domains is an alignment choice the adopting Dimension must declare explicitly, since the probability ranges carry assessment-specific conventions.", "Canonical statements are assumed to carry a BCP 47 language tag. Normalization behaviour for right-to-left scripts, non-Latin scripts and languages without whitespace word boundaries is left to the adopting Dimension, and the NFC rule may need supplementing there.", "Identifier namespace persistence assumes the owning package can maintain resolution for the full retention period; in jurisdictions or organisations without that capability, a governed external identifier authority should be used instead of local minting." ], "adversarial_checks": [ "Boundary test against WM-KNW-008: every bundle, layer, finding, data element and function was re-read for evidence description, source appraisal, citation formatting or evidence lifecycle content. Only binding parameters survive - target identifier, role, directionality, locally stated weight and binding status - and a dedicated validation question in the calibrated-confidence finding tests for appraisal grades leaking in.", "Enforcement and evaluation test: no node owns truth evaluation, verdict issuance, policy decisioning, validation execution or runtime audit trails. validate-claim-record was scoped to own shapes and profiles with execution delegated to an external processor; classify-and-express-access expresses ODRL policy with enforcement explicitly excluded, matching ODRL's own statement that it defines no enforcement mechanism; record-status-notice-reference records an externally issued notice rather than performing retraction; dispose-claim-record defines the rule while execution is owned by the adopting Dimension.", "Counterexample test on non-truth-apt input: questions, imperatives, performatives, bare exclamations and pure preferences were tested against the truth-aptness gate and are rejected with a reason code and a referral, so the register does not silently absorb non-propositional content under a 'claim' label.", "Identity test: date-based and text-based identity were both rejected. A content digest was demoted to a secondary equivalence key after considering two counterexamples - the same sentence expressing different claims in different contexts, and one claim restated across several texts and languages - which a digest-primary scheme would mishandle.", "Parent-duplication test against WM-KNW-001: generic knowledge-asset register mechanics, licensing and stewardship machinery were left to the CHILD link, and only claim-specific specialization (truth-aptness, qualification, assertion, claim relations) is modelled locally.", "Lossy-projection test: schema.org Claim, VC 2.0 and AIF projections were each checked for the ability to carry scope qualifiers, calibrated confidence and record status. Each loses at least one, so all three are declared lossy in the crosswalk and are prohibited as systems of record - which is why the crosswalk artifact carries per-profile lossiness rather than a single conformance statement.", "Single-finding-layer test: layers were checked for padding, and three candidate findings (assertion context, quality signals, exchange profile) were merged into adjacent findings rather than retained as thin nodes, so that no layer exists purely to inflate structure.", "Structural-support test: every node was checked for at least one primary source. The argumentation ownership node found no in-registry owner and is recorded as a gap in the checklist and as a placeholder composition link rather than being presented as canonical structure." ] }, "researchAdjudication": { "providerMode": "single-provider-waiver", "activeProviders": [ "claude" ], "waivedProviders": [ "grok" ], "providerPolicy": { "contract_version": "1.0.0", "mode": "single-provider-waiver", "effective_at": "2026-08-29T09:06:27Z", "scope": "Queued subject-model research from WM-XCT-013 onward", "active_providers": [ "claude" ], "waived_providers": [ { "provider": "grok", "authorized_by": "repository owner", "authorized_at": "2026-08-29T09:06:27Z", "reason": "The repository owner explicitly instructed the research queue to continue without Grok after repeated structured-output failures." } ], "review_rule": "Claude-only results require a separate no-tools adversarial audit and remain reviewable drafts with a visible single-provider hold." }, "boundaryDecision": { "entry_kind": "entity", "status": "accepted", "rationale": "The aggregate root is the claim record: primary identifier, version manifest, status transitions, tombstone, sensitivity class and access policy all attach to it, and no other node in the structure carries independent identity or lifecycle. The assertion act is correctly held inline in claimant-and-assertion-act as an event facet of the record rather than promoted to a sibling entry, and the register itself is the owning package (Dimension-owned), not the entry. The registry record's entry_kind 'standalone-mm' is a record-plane classification on a different axis from the research-plane 'entity' and is not a contradiction; both values must be carried side by side and the synthesizer must overwrite neither. Acceptance is conditioned on the identity-bearer wording amendment in decision 3, which does not change the classification." }, "decisions": [ { "concept": "Aggregate root is the claim record, with the assertion act as an inline facet", "disposition": "accepted", "rationale": "Identity, versioning, status, tombstone, sensitivity and access all attach to the record; the assertion act has no independent lifecycle here and would create a second identity for the same proposition if promoted to a sibling node." }, { "concept": "Research entry_kind 'entity' against frozen registry entry_kind 'standalone-mm'", "disposition": "accepted as non-conflicting (different classification axes)", "rationale": "The registry field classifies the record plane (standalone meta-model versus component), the research field classifies the subject. Treating these as a disagreement would manufacture a conflict; both values must survive synthesis unmodified." }, { "concept": "Conflict note declaring 'the proposition as the identity bearer' while restatements and translations are separate records", "disposition": "accepted with mandatory amendment", "rationale": "conflicts[0] names the proposition as identity bearer, but the equivalence finding and conflicts[2] deliberately allow several records per proposition (curator restatements are new records, not edits). Wording must narrow to 'the asserted statement as recorded', with co-propositional records linked by equivalence edges." }, { "concept": "Placeholder COMPOSITION link to the unregistered argumentation / inference-scheme owner", "disposition": "rejected", "rationale": "The frozen relationship contract carries only REFERENCE to WM-KNW-008, registry composition_role is REFERENCE and default_link_type is TYPED-EDGES. A composition link to an unregistered target contradicts the ledger the model itself declares authoritative; the gap must be carried as a registry gap entry only." }, { "concept": "CHILD link to WM-KNW-001 asserted in registry_links but absent from the frozen relation ledger extract", "disposition": "deferred to ledger reconciliation", "rationale": "parent_ids in the registry record supports the parent relation, but no CHILD row appears in the frozen contract while the model states the ledger is authoritative and that no local structure may contradict it. Either the row is added or the model must cite parent_ids as the link's source." }, { "concept": "serial_naming_rule enumerates only eight of the twelve artifacts marked serial:true", "disposition": "deferred — reconcile before publication", "rationale": "Claim Identifier Register, Claim Record Provenance Bundle, Claim Version Manifest and Claim Access Policy Expression are serial:true but omitted from the naming rule. Either extend the rule to all twelve or reclassify the append-only registers as continuous non-serial artifacts." }, { "concept": "Content digest granularity over append-only artifacts (provenance bundle, version manifest)", "disposition": "deferred — per-entry digest required", "rationale": "integrity_rule requires a digest over an artifact's canonical serialization and requires records to cite serial plus digest, but an append-only collection's digest changes on every append, so every citing record would fail integrity at the next entry. Digests must be per entry with an optional collection checkpoint." }, { "concept": "Append-only version manifest retaining prior statement text versus erasure-driven redaction of personal data", "disposition": "deferred — unresolved internal tension, published as a flagged draft issue", "rationale": "crud.delete permits redacting statement text under an erasure obligation while patch_rules and audit_requirements make manifests append-only and explicitly retain the prior text, so redaction can be defeated by the manifest. A manifest-redaction carve-out with a tombstoned entry is needed; this is amendable and does not block a reviewable draft." }, { "concept": "validate-claim-record function wording asserts execution the model disclaims", "disposition": "accepted with mandatory wording amendment", "rationale": "The model owns shapes and profiles and delegates execution to an external conformant processor, yet the function reads 'check a claim record ... producing a severity-graded conformance report', which is execution. Reword to commissioning validation against the published shapes and interpreting the returned report." }, { "concept": "dispose-claim-record function scope against Dimension-owned erasure execution", "disposition": "accepted with scope amendment", "rationale": "Writing the tombstone and redacting local statement text are in-scope register operations, but 'notifying dependents' and any implication of cross-store erasure encroach on the adopting Dimension's data-protection and storage platform, which delete rules explicitly reserve. Scope the function to the register's own record state plus a dangling-edge notice." }, { "concept": "Logical form / modality typology and the eight-way claim-type enumeration are uncited design proposals", "disposition": "accepted with mandatory addition to known_omissions", "rationale": "Polarity, negation scope, quantification, alethic/deontic modality and hedging are not defined by SRC-004, SRC-010 or SRC-005, and no cited source defines the empirical/causal/comparative/predictive/definitional/classificatory/normative/evaluative type list. known_omissions declares three design proposals but not these two, so the omissions list is materially incomplete." }, { "concept": "SRC-006 pinned to a GitHub master branch and SRC-014 flagged primary_source", "disposition": "deferred — repin and reclassify", "rationale": "Master-branch HTML is a moving target and must be pinned to a commit or release before it can support artifact-identity and canonicalization rules; the go-fair.org page is a derived restatement of the 2016 Scientific Data paper and should be marked secondary rather than primary." }, { "concept": "SRC-010 cited at the canonical ipcc.ch URL while the content was read from a mirror after HTTP 403", "disposition": "accepted with dual-locator citation required", "rationale": "The substitution is honestly declared in known_omissions, but a reviewer cannot reproduce the read from the cited URL. The citation must carry the canonical identifier and the actual retrieved access locator so the calibrated-uncertainty grounding stays verifiable." }, { "concept": "access.scopes omits the record scope its own exception depends on", "disposition": "deferred", "rationale": "scopes lists bundle, layer, finding and artifact, while the exceptions grant artifact-scope access independently of 'record-scope' access, which is undeclared. Add the record scope or restate the exception in the declared vocabulary so access rules stay machine-checkable." }, { "concept": "Single-finding layer validation-and-quality and single-layer bundle interoperability-and-agent-operation", "disposition": "accepted as-is, asymmetry recorded", "rationale": "Both nodes are thin but substantively distinct, and the adversarial single-finding-layer test addressed padding rather than thinness. With no second provider result there is no evidence base for inventing findings, and merging would bury the expression-versus-execution boundary the model depends on." }, { "concept": "Findings with no operational function: equivalence/restatement/translation, assessment-and-review-reference, argument-structure-role-reference", "disposition": "deferred — add_functions barred under the waiver", "rationale": "assert-claim-relation covers entailment, specialization, presupposition, conflict, preference and supersession but not equivalence, restatement or translation, and no function binds a review reference or an argument role. The gap is recorded as deferred research rather than silently patched, since single-provider mode forbids function additions." }, { "concept": "crud.create requirement for a 'truth-aptness verdict'", "disposition": "accepted with wording amendment", "rationale": "'Verdict' collides with the assert-what-is-asserted policy forbidding any field readable as a truth judgement. Rename to truth-aptness determination so the admissibility gate cannot be misread or projected as an evaluation of the proposition's truth." }, { "concept": "Question kind assignments 'composition' on ens-q2 and abe-q4", "disposition": "deferred — kind normalization", "rationale": "Specialization of a broader claim is a taxonomic relation and consulting referenced models is a dependency; neither is part-whole composition. The kind histogram feeds downstream coverage analytics, so mislabelling distorts it without changing question substance." }, { "concept": "No registered legacy or previous-version material", "disposition": "accepted", "rationale": "The frozen legacy source is empty, so no prior-art reconciliation is possible or required. The published draft must state that no legacy baseline exists rather than implying continuity with earlier Vercy material." }, { "concept": "Tombstone persistence horizon after the retention period expires", "disposition": "deferred", "rationale": "The access exception requires retracted, withdrawn and superseded claims to stay readable so downstream copies remain reconcilable, but retention is bound to sensitivity class and terminal status with no floor for the tombstone itself, so a disposed record's tombstone may lawfully vanish and break reconciliation." } ], "publicationHolds": [ "Live source and version verification hold: before publication, re-resolve all fourteen source URLs and confirm each version pin — schema.org 30.0 (2026-03-19) for SRC-002/003, RDF 1.2 Concepts CR Snapshot 07 April 2026 (SRC-009, still a moving target), VC 2.0 REC 15 May 2025, PROV-O, SHACL, ODRL 2.2, RFC 3339, NISO RP-45-2024, JCGM 200:2012 and the AIF PDF — repin SRC-006 from master branch to a commit or release, reclassify SRC-014 as a derived restatement, and record the SRC-010 mirror as the actual access locator alongside the canonical ipcc.ch URL.", "Single-provider hold, owner-authorized: independent second-provider review is absent because the repository owner waived Grok on 2026-08-29T09:06:27Z after repeated structured-output failures. This result rests on one provider plus this no-tools adversarial audit, entry-kind agreement is 'waived' with no cross-provider corroboration of any bundle, layer, finding or source, and every publication artifact must display the waiver and keep the model a reviewable draft rather than a canonical record.", "Registry and relationship-contract hold: review_state stays boundary-review-required until the relation ledger carries an explicit WM-KNW-001 CHILD row (or the model cites parent_ids as the link's source), the placeholder composition link to the unregistered argumentation owner is removed, and the argumentation / inference-scheme ownership gap is formally raised against the registry.", "Artifact identity hold: reconcile serial_naming_rule with all twelve artifacts marked serial:true and define per-entry digests for the append-only provenance bundle and version manifest before any claim record is allowed to cite artifact serial plus digest as a promotion precondition.", "Privacy and erasure hold: the redaction rule and the append-only version manifest are in unresolved tension, and the GDPR basis remains an uncited regional assumption because EUR-Lex retrieval was blocked; the adopting Dimension's counsel must confirm the accuracy, rectification and erasure duties before any record containing personal data is promoted out of draft.", "Uncited-typology hold: known_omissions must be extended to declare the logical-form/modality typology and the eight-way claim-type enumeration as design proposals before publication, so readers do not read them as standards-backed vocabularies.", "Independent second-provider review was explicitly waived by the repository owner; this Claude-only result remains a reviewable draft." ], "deferredResearch": [ "Register an owner for argument-scheme definition and for the application and evaluation of inference, conflict and preference rules, then replace the AIF I-node role-reference gap with a real in-registry target and delete the placeholder composition link.", "Obtain a retrievable primary legal citation (EUR-Lex Regulation (EU) 2016/679 or equivalent) for accuracy, rectification and erasure duties so the privacy dimension moves from regional assumption to cited requirement, and resolve how erasure interacts with append-only version manifests.", "Verify JCGM 200:2012 VIM terms (measurand, measurement result, measurement uncertainty, coverage interval, coverage probability, metrological traceability) from full text rather than the BIPM catalogue description, and keep the metrology link declared as alignment until that succeeds.", "Search for, or formally propose and publish, a normative registry of claim-to-claim relation types; the current entailment/specialization/presupposition/conflict/preference/supersession codes are a design synthesis of AIF conflict and preference semantics with PROV revision semantics.", "Add operational functions in the next provider pass for equivalence/restatement/translation linking, assessment and review reference binding, and argument-structure role binding, all of which are findings without functions today.", "Determine normalization behaviour beyond NFC for right-to-left scripts, non-Latin scripts and languages without whitespace word boundaries, and define whether cross-language equivalence thresholds stay entirely Dimension-owned.", "Decide whether deontic and normative claim content needs a dedicated model rather than the single modality field, including the relation between a norm asserted as claim content and ODRL policy expressed about the record." ] }, "statistics": { "sources": 14, "bundles": 6, "layers": 12, "findings": 28, "questions": 112, "artifacts": 12, "functions": 13 } }