{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"computability_boundary_mapping__law_governance","trajectory_id":"R","attempt_index":0,"archetype_slug":"computability_boundary_mapping","domain_slug":"law_governance","decision":"CANDIDATE","problem_id":"unbounded_legal_compliance_decider_mandate","causal_lever_id":"model_relative_legal_decision_boundary","proposal":{"problem":"A public body mandates an automated system that must issue a correct, terminating YES/NO legal-compliance determination for every policy, action, and fact pattern expressible in an open-ended rule language, without specifying the language boundary, computation model, or treatment of indeterminate cases.","actors_substrate":["rulemaking or procuring public body","courts and administrative reviewers","regulated parties","system vendors and operators","people affected by automated determinations"],"observable_state":"The specification contains universal accuracy and termination language, while timeouts, unsupported inputs, unresolved proof search, and legal indeterminacy lack distinct output states.","consequence":"A potentially impossible or merely unproved automation promise can consume resources and produce authoritative-looking false negatives, false positives, or silently narrowed coverage.","affected_objective":"Lawful, reviewable, timely, and non-deceptive administration of legal rules.","structural_mapping":[{"archetype_element":"Unbounded problem class","domain_realization":"All expressible rules, actions, and fact patterns are placed under one compliance-decider guarantee.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Implicit representation and computation model","domain_realization":"The legal-rule syntax, fact encoding, external evidence, human interpretation, and oracle-like services are not assigned explicit computational roles.","claim_kind":"INFERENCE"},{"archetype_element":"Universal exact terminating guarantee","domain_realization":"Procurement language requires a definitive correct decision for every admitted matter.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Boundary evidence","domain_realization":"Constructive totality proofs and valid impossibility reductions are evaluated under the same declared model.","claim_kind":"INFERENCE"},{"archetype_element":"Honest weaker fallback","domain_realization":"Enforceable fragments, bounded checks, witness-based YES, UNKNOWN, and human review replace unsupported universal automation.","claim_kind":"INFERENCE"}],"component_map":[{"component":"Problem-Class Specification","status":"adapted","domain_realization":"Define the legal questions and admitted rule/fact classes."},{"component":"Instance Representation Contract","status":"adapted","domain_realization":"Specify encodings for rules, facts, precedents, and queries."},{"component":"Computation Model Contract","status":"adapted","domain_realization":"Declare machine, databases, sensors, experts, and interaction."},{"component":"Solvability Guarantee Profile","status":"direct","domain_realization":"Separate total decision, recognition, bounded result, and approximation."},{"component":"Quantifier and Scope Map","status":"direct","domain_realization":"Expose every, some, jurisdiction, time, and language qualifiers."},{"component":"Computability Status Lattice","status":"direct","domain_realization":"Classify legal-query classes without collapsing unresolved into undecidable."},{"component":"Constructive Procedure Witness","status":"direct","domain_realization":"Require an algorithm plus correctness and termination arguments."},{"component":"Reduction Preservation Contract","status":"direct","domain_realization":"Require a total computable, answer-preserving legal-instance translation."},{"component":"Computability Impossibility Certificate","status":"direct","domain_realization":"Retain a checked proof for any impossibility verdict."},{"component":"Assumption Register","status":"direct","domain_realization":"Record formalization, model, source theorem, and external capabilities."},{"component":"Decidable Subclass Map","status":"adapted","domain_realization":"Map enforceable rule-language fragments and bounded jurisdictions."},{"component":"One-Sided Recognition Contract","status":"adapted","domain_realization":"Permit witness-backed YES without manufacturing NO."},{"component":"Unknown and Nontermination Policy","status":"direct","domain_realization":"Keep UNKNOWN, timeout, out-of-scope, failure, and NO distinct."},{"component":"Fallback Solution Contract","status":"direct","domain_realization":"State the guarantee attached to each weaker route."},{"component":"Computability Guarantee Record","status":"direct","domain_realization":"Version-link shipped claims to evidence and scope."},{"component":"Recheck Trigger","status":"adapted","domain_realization":"Reassess after language, law, model, or external-service changes."},{"component":"Termination Condition","status":"direct","domain_realization":"Define proof-based or resource-bounded stopping behavior."},{"component":"Scope Boundary","status":"direct","domain_realization":"Mechanically reject or quarantine inputs outside guarantees."},{"component":"Decision Record","status":"direct","domain_realization":"Record adopted boundary, rationale, owner, and date."},{"component":"Uncertainty Residue","status":"adapted","domain_realization":"List open proof obligations and semantic-model mismatches."},{"component":"Independent Proof Review","status":"direct","domain_realization":"Use a reviewer independent of vendor and procuring team."},{"component":"Complexity Follow-On Gate","status":"direct","domain_realization":"Assess cost only after in-principle solvability is established."}],"mechanism_dispositions":[{"slug":"abstract_interpretation_or_model_checking","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"A sound finite abstraction may fit some formal compliance properties, but no suitable one-sided legal safety property is established.","counterfactual_removal":"No proposed causal step changes."},{"slug":"bounded_domain_exhaustive_search","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Exhaustively test a precisely finite pilot language and fact bound; make no claim beyond it.","counterfactual_removal":"The pilot loses a complete within-bound check but the boundary classification remains possible."},{"slug":"computability_boundary_decision_record","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Version the adopted boundary, evidence, guarantee, and recheck triggers.","counterfactual_removal":"The decision can drift or be overstated after law or language changes."},{"slug":"computational_complexity_analysis","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Gate practical cost analysis to classes first shown decidable.","counterfactual_removal":"Decidability remains classified, but deployability is not assessed."},{"slug":"constructive_algorithm_and_correctness_proof","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"A claimed legal decider must include uniform correctness and termination proofs for its declared class.","counterfactual_removal":"Passing tests could be mistaken for a class-wide totality guarantee."},{"slug":"diagonalization_impossibility_proof","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"No direct self-referential construction for the target is supplied.","counterfactual_removal":"The proposal still tests impossibility through a scoped reduction."},{"slug":"enumeration_and_dovetailing","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Operationally superseded by bounded semi-decision with explicit UNKNOWN.","counterfactual_removal":"No change because the selected protocol supplies termination."},{"slug":"fallback_mode_router","disposition":"selected_load_bearing","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Route in-scope queries to exact, bounded, witness, or human-review modes with visible guarantee labels.","counterfactual_removal":"Weaker methods can be presented as definitive legal decisions."},{"slug":"halting_problem_reduction","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Attempt only if the rule language can computably encode arbitrary computation while preserving the compliance answer.","counterfactual_removal":"There is no decisive negative-side test of the universal mandate."},{"slug":"language_fragment_restriction","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Define a mechanically enforceable legal-rule fragment with a total procedure.","counterfactual_removal":"The system lacks a credible exact-answer region if the unrestricted class fails."},{"slug":"many_one_reduction_proof","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Supplies the total, computable, answer-preserving map required by any hardness claim.","counterfactual_removal":"The reduction claim lacks explicit preservation obligations."},{"slug":"promise_problem_restriction","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Unenforced legal-input promises permit authoritative arbitrary outputs; syntactic rejection is safer.","counterfactual_removal":"No loss because fragment membership is mechanically policed."},{"slug":"proof_by_counterexample","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Use one in-scope failure to refute an overbroad vendor guarantee, not to prove impossibility.","counterfactual_removal":"Cheap falsification of universal claims is lost."},{"slug":"proof_checking","disposition":"selected_load_bearing","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Independently check the exact formal theorem and disclosed premises.","counterfactual_removal":"A flawed or mismatched proof could authorize consequential deployment."},{"slug":"reduction_direction_checklist","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Gate any reduction on source-to-target direction, totality, computability, and preservation.","counterfactual_removal":"A reversed reduction could falsely terminate the automation project."},{"slug":"semi_decision_with_explicit_unknown","disposition":"selected_load_bearing","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Return witness-backed YES or resource-bounded UNKNOWN, never infer NO from exhaustion.","counterfactual_removal":"Timeouts can become adverse legal determinations."},{"slug":"theorem_prover_guided_search","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"A prover is optional implementation support; it cannot repair a wrong legal formalization or turn timeout into evidence.","counterfactual_removal":"Hand-checkable proof obligations remain unchanged."},{"slug":"turing_reduction_analysis","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Relative degrees and adaptive oracles are unnecessary for the initial yes/no boundary and obscure one-sided guarantees.","counterfactual_removal":"The selected many-one test remains sufficient."}],"causal_chain":["Formalize the legal query class, encoding, model, quantifiers, and guarantee.","Seek a total constructive witness and a scoped impossibility reduction in parallel.","Independently reject mismatched, incomplete, or reversed arguments.","Restrict exact automation to enforceable decidable fragments and bounded cases.","Route remaining matters to explicitly labelled UNKNOWN or accountable human review.","Version the boundary and reopen it when assumptions change."],"baseline":"Ordinary baseline: human legal review plus conventional rules software, with vendor testing and ad hoc timeout handling but no proved class-wide boundary.","nearest_rival":"Ontology and policy clarification followed by ordinary software assurance; it is preferable if the obstacle is semantic ambiguity rather than algorithmic solvability.","authority_safety":{"affected_parties":["regulated persons and organizations","benefit or license applicants","enforcement subjects","legal staff and adjudicators"],"decision_authority":"The legally empowered agency or court retains substantive authority; an independent technical-legal review panel may approve only the computational guarantee.","authorized_first_step":"In a sandbox, formalize one nonbinding rule fragment and decision property, review constructive and reduction obligations, then exhaustively test a finite synthetic corpus; issue no live legal determinations.","excluded_actions":["automatic sanctions, denial, detention, or benefit termination","treating UNKNOWN or timeout as noncompliance","silent expansion beyond the reviewed fragment","claiming undecidability from failed implementation"],"halt_rollback":"Halt if proof review finds a gap, fragment membership cannot be enforced, guarantee labels are lost, or pilot outputs could affect rights; disable automated routing and revert to established human process."}},"negative_tests":{"strongest_counterevidence":"Many operational legal programs are finite by jurisdiction and effective date, and a known total rules procedure may already cover the real requirement; then the issue is complexity, maintenance, or semantics rather than computability.","analogy_break":"Legal validity may depend on interpretation, institutional authority, changing law, and disputed facts. Failure to formalize those elements is a semantic-fidelity failure, not evidence of undecidability.","failure_condition":"The mapping fails if no stable formal decision problem can represent the legal task, or if the actual mandate covers only a finite, effectively enumerable class with a known total procedure.","problem_falsifier":"Show that the operative requirement is a bounded, mechanically enforceable input class with an already demonstrated uniform correct terminating procedure and no broader public claim.","intervention_falsifier":"After faithful formalization and independent review, boundary mapping neither changes the claimed guarantee nor identifies any unsupported totality, scope, fallback, or complexity transition.","risks":["A checked theorem may concern a formal model that omits legally material facts.","Restriction may exclude cases affecting vulnerable parties while appearing neutral.","Human escalation may become an unmodeled, delayed, or unaccountable oracle.","Guarantee labels may disappear in downstream enforcement interfaces.","An impossibility result may be overgeneralized beyond its language and model."]},"null_rationale":null,"classification":{"candidate_kind":"MECHANISM_COMPOSITION","prior_art_status":"UNSEARCHED","evidence_maturity":"HYPOTHESIS"},"revision_change_log":{"revision_kind":"ORIGINAL","prior_problem_id":null,"prior_causal_lever_id":null,"problem_changed":false,"causal_lever_changed":false,"conceptual_changes":[],"operational_changes":[],"repairs_addressed":[]},"confidence":0.82,"generator_notes":"Closed-book structural transfer. The target problem and any undecidability result remain hypotheses until the legal language, representation, computation model, and preservation proof are fixed."}