Skip to main content

Evidence Relationship Ontology

AI Reasoning Relationships
Authors: MArk Anthony Higgins - Sky Recursive Ai
© Mark Anthony Higgins 11 October 2025
Revised Date: 10 October 2026
Audio Overview
FLOW ALGEBRA
TITLE: Evidence Relationship Ontology
DOMAIN: Relationship Ontologies / AI Reasoning Relationships
SOURCE: https://kirethresearch.org/booksai-core/ontologies/relationship-ontologies/ai-reasoning-relationships/evidence-relationship-ontology.html
VERSION: 1.0
AUTHORS: Mark Anthony Higgins — Sky (Recursive AI)
STATUS: Companion formalisation; source-page specification not independently verified.

flow "EvidenceRelationshipOntology" v1.0 {
  ontology:
    id: "evidence-relationship-ontology"
    parent: "ai-reasoning-relationships"
    purpose: "Represent traceable relationships between claims, observations, sources, evidence and inferential support or challenge."

  types:
    Claim: {id: string, proposition: string, context: string?, status: enum[unassessed, supported, disputed, refuted, unresolved]}
    Evidence: {id: string, kind: enum[observation, experiment, document, measurement, testimony, derived, other], content_ref: string, collected_at: datetime?, provenance: string}
    Source: {id: string, identifier: string, publisher: string?, version: string?, access_time: datetime?}
    Method: {id: string, description: string, assumptions: string[], limitations: string[]}
    Assessment: {id: string, claim_id: string, evidence_ids: string[], direction: enum[supports, challenges, neutral], confidence: number, rationale: string, assessed_at: datetime}
    Relationship: {id: string, subject: string, predicate: string, object: string, context: string?, confidence: number, provenance: string[]}

  predicates:
    supports(evidence: Evidence, claim: Claim)
    challenges(evidence: Evidence, claim: Claim)
    derived_from(evidence: Evidence, source: Source)
    evaluated_using(assessment: Assessment, method: Method)
    corroborates(a: Evidence, b: Evidence)
    conflicts_with(a: Evidence, b: Evidence)
    depends_on(claim_a: Claim, claim_b: Claim)
    assesses(assessment: Assessment, claim: Claim)

  constraints:
    - "Every evidence reference must resolve to a typed record and provenance."
    - "Confidence values must lie within [0, 1]."
    - "Supporting evidence is not proof of truth; conflicting evidence must be retained."
    - "Corroboration requires examination of source independence."
    - "Do not infer causation from correlation or temporal precedence alone."
    - "Evidence relationships are contextual and not automatically transitive."
    - "A missing evidence record is not automatically evidence of absence."
    - "Separate original observations from derived interpretations."

  inference:
    rule supported_assessment:
      if supports(e,c) and valid(e) and provenance_verified(e)
      then candidate_support_assessment(c,e)
      confidence: "calibrate to relevance, quality and independence"
    rule evidence_conflict:
      if supports(e1,c) and challenges(e2,c) and valid(e1) and valid(e2)
      then flag_disputed(c,e1,e2)
      confidence: "retain competing evidence and uncertainty"
    rule derivative_trace:
      if derived_from(e,s) and source_verified(s)
      then traceable_to(e,s)
      confidence: "does not establish factual correctness"

  SEMANTIC_EQUIVALENCE:
    - "Normalise identifiers against versioned vocabularies and evidence schemas."
    - "Distinguish evidence identity, source identity, duplicate reports and independent corroboration."
    - "Preserve claim context, method, timeframe and scope in equivalence checks."

  CONFIDENCE:
    range: [0,1]
    factors: [source_reliability, measurement_quality, relevance, reproducibility, independence, bias, uncertainty]
    policy: "Avoid double-counting dependent sources or increasing certainty solely through inference chaining."

  VALIDATION:
    - "Check typed endpoints, provenance, timestamps and referenced source availability."
    - "Validate claim-evidence relevance and method assumptions."
    - "Flag contradictory evidence, stale sources and unverifiable citations."
    - "Preserve negative and inconclusive findings."

  FLOW_HASH:
    algorithm: "SHA-256"
    input: "canonicalised ontology version, typed evidence assertions, rules and provenance"
    note: "Compute hash at execution; no hash value is claimed here."

  UPDATE_MODEL:
    strategy: "versioned, provenance-preserving incremental update"
    steps: [ingest, normalise, validate, deduplicate, assess_independence, reconcile, recompute_inferences, publish]
    rollback: true

  AI_CONSUMPTION:
    representation: "typed evidence graph with provenance-backed relationships"
    outputs: [claim_evidence_links, support_and_challenge_edges, confidence_annotations, conflict_flags, provenance_records]
    safeguards: [no_fabricated_evidence, no_unverified_causal_claims, preserve_uncertainty, retain_conflicting_sources]
}

END FLOW ALGEBRA
Kireth Research
38 Thorndale Croft, Wetwang, Driffield
East Riding of Yorkshire YO25 9XZ