Legal Reasoning Domain Ontology

FLOW ALGEBRA
TITLE: Legal Reasoning Domain Ontology
VERSION: 1.0
DOMAIN: Legal and Government Domains / Legal Reasoning
SOURCE_URL: https://kirethresearch.org/booksai-core/ontologies/domain-ontologies/legal-government-domains/legal-reasoning-domain-ontology.html
STATUS: Domain-based ontology draft; source page not independently verified.
AUTHORS: Mark Higgins — Sky (Recursive AI)

FLOW "LegalReasoningDomainOntology" v1.0

PURPOSE:
  Represent and reason over legal norms, facts, evidence, precedents,
  interpretations and arguments with explicit jurisdiction, temporal scope,
  uncertainty, provenance and human-review boundaries.

TYPES:
  Jurisdiction: {id, name, level, valid_period}
  LegalSource: {id, source_type, citation, jurisdiction, effective_period, authority_rank}
  LegalNorm: {id, source_id, condition, consequence, exceptions, status}
  LegalConcept: {id, label, definition, jurisdiction}
  LegalFact: {id, proposition, event_time, evidence_ids, disputed}
  Evidence: {id, kind, origin, collected_at, reliability, admissibility_status}
  LegalIssue: {id, question, relevant_facts, relevant_norms}
  LegalArgument: {id, claim, premises, rule_ids, counterarguments}
  Precedent: {id, court, holding, material_facts, jurisdiction, decision_date}
  Interpretation: {id, norm_id, method, reading, supporting_sources}
  LegalConclusion: {id, issue_id, conclusion, reasoning_trace, confidence, review_status}
  LegalActor: {id, role, jurisdiction}
  Proceeding: {id, forum, stage, parties, issues}

RELATIONSHIPS:
  APPLIES_IN(LegalNorm, Jurisdiction)
  DERIVES_FROM(LegalNorm, LegalSource)
  SUPPORTS(Evidence, LegalFact)
  CONTRADICTS(Evidence|LegalFact|LegalArgument, Evidence|LegalFact|LegalArgument)
  RAISES(LegalFact, LegalIssue)
  GOVERNS(LegalNorm, LegalIssue)
  INTERPRETS(Interpretation, LegalNorm)
  CITES(LegalArgument, LegalSource|Precedent)
  DISTINGUISHES(Precedent, Precedent)
  OVERRULES(Precedent, Precedent)
  SUPPORTS_CONCLUSION(LegalArgument, LegalConclusion)
  SUBJECT_TO_EXCEPTION(LegalNorm, LegalNorm)
  REVIEWED_BY(LegalConclusion, LegalActor)

INFERENCE_RULES:
  R1 SOURCE_VALIDITY:
    IF source.jurisdiction matches issue.jurisdiction
       AND issue.event_time within source.effective_period
       AND source.status is operative
    THEN source is a candidate applicable authority.
  R2 NORM_APPLICATION:
    IF norm is applicable AND all necessary conditions are established
       AND no applicable exception defeats the norm
    THEN derive a provisional consequence with a trace to premises.
  R3 PRECEDENT_ANALOGY:
    IF precedent is authoritative in the jurisdiction
       AND legally material facts align
    THEN add precedent as support; record distinctions explicitly.
  R4 CONFLICT_RESOLUTION:
    IF candidate norms conflict
    THEN evaluate authority, specificity, temporal priority,
         jurisdiction and recognised interpretative rules;
         preserve unresolved conflicts for human review.
  R5 EVIDENCE_ASSESSMENT:
    IF fact is disputed OR evidence is inadmissible/uncertain
    THEN mark the dependent conclusion as contested or conditional.
  R6 DEFEASIBLE_REASONING:
    IF new evidence, an exception, or controlling authority defeats a premise
    THEN retract or revise dependent provisional conclusions.
  R7 ARGUMENT_TRACEABILITY:
    FOR EACH conclusion preserve source citations, applied rules,
    facts, assumptions, objections and alternative interpretations.

CONFIDENCE:
  type: calibrated_score[0,1] | qualitative_uncertainty
  factors: [source_authenticity, source_authority, factual_support,
            interpretative_ambiguity, precedent_fit, temporal_validity]
  rule: confidence is epistemic uncertainty, not a probability of court success.

SEMANTIC_EQUIVALENCE:
  compare: [concept_definition, jurisdiction, effective_period, legal_context]
  rule: matching labels alone do not establish equivalent legal meaning.
  aliases: ["legal inference", "juridical reasoning", "legal argumentation"]

PROVENANCE:
  required: [source_id, citation, retrieval_or_record_date, jurisdiction,
             effective_date, evidence_origin, transformation_history]
  rule: separate asserted facts, source quotations, interpretations and conclusions.

FLOW_HASH:
  method: SHA-256
  input: canonical_serialisation(types, relationships, rules, version, source_refs)
  rule: recompute after any semantic modification; no hash value is asserted here.

UPDATE_MODEL:
  triggers: [legislation_amended, precedent_overruled, new_evidence,
             jurisdiction_changed, correction, interpretation_revised]
  steps: [validate_change, version_sources, identify_dependants,
          recompute_conclusions, update_confidence, log_changes, human_review]

AI_CONSUMPTION:
  output: [typed_entities, grounded_relationships, applicable_sources,
           reasoning_trace, competing_arguments, uncertainty, provenance,
           review_required, version_metadata]
  safeguards: [no_fabricated_authorities, jurisdiction_scoping,
               temporal_scoping, distinguish_law_from_opinion,
               preserve_counterarguments, human_legal_review]

VALIDATION:
  require_unique_ids: true
  require_citations_for_authorities: true
  reject_unscoped_legal_claims: true
  flag_conflicting_authorities: true
  flag_missing_effective_dates: true
  flag_unreviewed_high_stakes_conclusions: true

END FLOW