{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"computability_boundary_mapping__chemistry_materials","trajectory_id":"R","attempt_index":0,"archetype_slug":"computability_boundary_mapping","domain_slug":"chemistry_materials","decision":"CANDIDATE","problem_id":"universal_exact_reaction_network_hazard_reachability","causal_lever_id":"enforceable_scope_and_labeled_fallback_routing","proposal":{"problem":"A materials or process-safety project claims an exact, always-terminating analyzer that decides whether any finitely described, potentially unbounded reaction network will ever produce a designated hazardous species. Timeouts and unsuccessful simulations are currently liable to be reported as safe, although the unrestricted network language may encode behavior for which no total decider exists.","actors_substrate":["computational chemists and materials modelers","process-safety engineers","reaction-rule models and their simulators","operators, downstream users, and nearby communities"],"observable_state":"Unrestricted hazard-reachability claims coexist with simulation timeouts, implicit model bounds, and outputs that collapse unknown into safe.","consequence":"A false safe verdict can expose affected parties; an unsupported impossibility claim can also abandon useful analyzers for enforceable fragments.","affected_objective":"Truthful and operationally useful hazard screening with no silent loss of safety guarantees.","structural_mapping":[{"archetype_element":"Open-ended problem class","domain_realization":"All networks expressible in a declared reaction-rule language, rather than one bounded reactor model.","claim_kind":"INFERENCE"},{"archetype_element":"Universal exact terminating answer","domain_realization":"For every valid network, return whether hazardous species is ever reachable and always halt.","claim_kind":"INFERENCE"},{"archetype_element":"Potential computability boundary","domain_realization":"The unrestricted language may simulate unbounded computation, making universal reachability undecidable.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Decidable region","domain_realization":"Finite-state, bounded-population, bounded-depth, or syntactically restricted network classes.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Honest weaker fallback","domain_realization":"Sound over-approximation, bounded exhaustive search, or witness-confirming analysis with explicit UNKNOWN.","claim_kind":"INFERENCE"},{"archetype_element":"Reclassification trigger","domain_realization":"Changes to reaction syntax, population bounds, kinetics, external observations, or claimed guarantee.","claim_kind":"INFERENCE"}],"component_map":[{"component":"Problem-Class Specification","status":"adapted","domain_realization":"Define network language, hazard predicate, initial states, and permitted unbounded behavior."},{"component":"Instance Representation Contract","status":"adapted","domain_realization":"Version the species, reactions, parameters, bounds, and encoding validity rules."},{"component":"Computation Model Contract","status":"adapted","domain_realization":"State digital procedure, arithmetic semantics, randomness, sensors, and external expert access."},{"component":"Solvability Guarantee Profile","status":"adapted","domain_realization":"Separate total exact decisions, sound alarms, bounded results, and witness-only confirmations."},{"component":"Quantifier and Scope Map","status":"direct","domain_realization":"Distinguish every encoded network from named models and bounded subclasses."},{"component":"Computability Status Lattice","status":"adapted","domain_realization":"Classify each scope as decidable, recognizable, relative, unresolved, or hypothesized undecidable."},{"component":"Constructive Procedure Witness","status":"adapted","domain_realization":"Require algorithm, correctness argument, and termination argument for any decidable claim."},{"component":"Reduction Preservation Contract","status":"adapted","domain_realization":"Require a total computable encoding preserving machine halting iff chemical hazard reachability."},{"component":"Computability Impossibility Certificate","status":"adapted","domain_realization":"Accept undecidability only after a checked reduction matching the deployed language."},{"component":"Assumption Register","status":"direct","domain_realization":"Record realizability, encoding, bounds, numerical semantics, and oracle assumptions."},{"component":"Decidable Subclass Map","status":"adapted","domain_realization":"List enforceable fragments and their lost chemical expressiveness."},{"component":"One-Sided Recognition Contract","status":"adapted","domain_realization":"A verified hazard witness may yield YES; lack of one never yields SAFE."},{"component":"Unknown and Nontermination Policy","status":"direct","domain_realization":"Keep UNKNOWN, TIMEOUT, OUT_OF_SCOPE, UNSAFE, and SAFE distinct."},{"component":"Fallback Solution Contract","status":"adapted","domain_realization":"Label every fallback by scope, direction of soundness, and resource bound."},{"component":"Computability Guarantee Record","status":"direct","domain_realization":"Version-link claims, proofs, implementations, and user-facing labels."},{"component":"Recheck Trigger","status":"adapted","domain_realization":"Reassess after language, model, bound, oracle, or guarantee changes."},{"component":"Termination Condition","status":"adapted","domain_realization":"Bound fallback time, depth, states, or molecule counts; return UNKNOWN on exhaustion."},{"component":"Scope Boundary","status":"adapted","domain_realization":"Mechanically reject or quarantine models outside guaranteed fragments."},{"component":"Decision Record","status":"direct","domain_realization":"Record the chosen boundary, evidence, fallback, owner, and expiry conditions."},{"component":"Uncertainty Residue","status":"adapted","domain_realization":"Retain unproved reduction obligations and abstraction/model mismatch."},{"component":"Independent Proof Review","status":"direct","domain_realization":"A reviewer checks theorem statement, reduction direction, and formalization fidelity."},{"component":"Complexity Follow-On Gate","status":"adapted","domain_realization":"Only decidable fragments proceed to scaling and deployment-feasibility analysis."}],"mechanism_dispositions":[{"slug":"abstract_interpretation_or_model_checking","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Use finite over-approximations only when their soundness direction is documented.","counterfactual_removal":"Fallback remains possible, but loses a potentially sound hazard-screening mode."},{"slug":"bounded_domain_exhaustive_search","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Exhaust networks or states only inside explicit finite bounds.","counterfactual_removal":"The proposal loses its exact, terminating bounded mode."},{"slug":"computability_boundary_decision_record","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Version the boundary, guarantees, evidence, and recheck triggers.","counterfactual_removal":"Guarantees can drift without an auditable decision trail."},{"slug":"computational_complexity_analysis","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Assess cost only after a fragment is classified decidable.","counterfactual_removal":"Computability remains classified, but deployability is not assessed."},{"slug":"constructive_algorithm_and_correctness_proof","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Require totality and correctness evidence for positive decidability claims.","counterfactual_removal":"A successful simulator could be mistaken for a class-wide decider."},{"slug":"diagonalization_impossibility_proof","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"A chemistry-specific direct self-reference construction is not supplied; reduction is the clearer test.","counterfactual_removal":"No change; the selected reduction route remains."},{"slug":"enumeration_and_dovetailing","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Fair enumeration is unnecessary for the bounded first test; a domain witness checker can supply YES evidence.","counterfactual_removal":"No material change to the bounded protocol."},{"slug":"fallback_mode_router","disposition":"selected_load_bearing","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Route by enforceable scope to exact, sound-approximate, bounded, or escalated modes.","counterfactual_removal":"Weaker results could be presented under the universal guarantee."},{"slug":"halting_problem_reduction","disposition":"selected_load_bearing","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Attempt a source-to-target encoding preserving halting as hazard reachability.","counterfactual_removal":"The hypothesized impossibility lacks a decisive evaluation path."},{"slug":"language_fragment_restriction","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Define mechanically checkable reaction-language fragments with known procedures.","counterfactual_removal":"There is no enforceable route from unrestricted claims to total guarantees."},{"slug":"many_one_reduction_proof","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Its obligations are incorporated into the selected specialized halting reduction.","counterfactual_removal":"No change if the specialized reduction discharges totality and preservation."},{"slug":"promise_problem_restriction","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Unenforced chemical promises could authorize plausible wrong answers; syntactic fragments are safer.","counterfactual_removal":"No change to the enforceable scope design."},{"slug":"proof_by_counterexample","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Use in-scope networks to refute overbroad analyzer claims, never to prove undecidability.","counterfactual_removal":"Universal overclaims become harder to falsify cheaply."},{"slug":"proof_checking","disposition":"selected_load_bearing","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Independently check constructive or impossibility certificates and their exact theorem statements.","counterfactual_removal":"A proof gap could incorrectly clear or prohibit a safety analyzer."},{"slug":"reduction_direction_checklist","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Gate the reduction on source-to-target direction, totality, and assumptions.","counterfactual_removal":"A common direction error is less likely to be caught early."},{"slug":"semi_decision_with_explicit_unknown","disposition":"selected_load_bearing","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"At a resource bound, return UNKNOWN rather than SAFE when no hazard witness appears.","counterfactual_removal":"Timeout can again become a false safety verdict."},{"slug":"theorem_prover_guided_search","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Optional automation is not required for the first hand-checkable boundary test.","counterfactual_removal":"No causal change; independent proof checking remains."},{"slug":"turing_reduction_analysis","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Relative degrees and adaptive oracles exceed the binary boundary question posed here.","counterfactual_removal":"No change to the proposed classification."}],"causal_chain":["Specify the reaction-network class, encoding, computation model, quantifiers, and hazard predicate.","Seek both a total constructive procedure and a valid impossibility reduction.","Independently review either certificate and retain unresolved obligations.","Enforce decidable fragments and finite bounds at input validation.","Route accepted queries to the strongest warranted mode and label its guarantee.","Return UNKNOWN on exhausted one-sided or bounded analysis, preventing timeout from becoming SAFE.","Send decidable scopes to complexity assessment and recheck the boundary when assumptions change."],"baseline":"Run numerical or symbolic simulation until success or timeout, then treat observed hazard as unsafe and prolonged non-observation as safe or failed.","nearest_rival":"Improve simulator performance and scaling while retaining one unrestricted Boolean hazard claim; this addresses complexity but not whether a total exact procedure exists.","authority_safety":{"affected_parties":["laboratory and plant workers","nearby communities","operators and downstream decision makers","model authors whose work may be excluded"],"decision_authority":"The process-safety owner and scientific-methods lead jointly approve labels and scope; an independent reviewer approves computability claims.","authorized_first_step":"For one versioned reaction language and one hazard predicate, conduct a four-week desk test: formalize the class, attempt the reduction and a constructive bounded solver, independently review both, and shadow-route 20 archived models without changing physical operations.","excluded_actions":["Declaring an unrestricted network class undecidable before a valid certificate","Reporting timeout, no witness, or out-of-scope as SAFE","Changing reactor operation or releasing materials during the desk test","Using an unenforced promise or hidden input restriction"],"halt_rollback":"Halt the pilot on any mislabeled guarantee, unsound abstraction, invalid reduction step, or scope-check failure; withdraw affected records and revert to existing human-reviewed safety procedures."}},"negative_tests":{"strongest_counterevidence":"Industrially relevant reaction models may all have enforced finite populations, horizons, or state spaces; then exhaustive decision is possible in principle and the real problem is only complexity.","analogy_break":"A formal reaction-rule network is not automatically a faithful physical chemical system. A reduction may depend on unbounded counts, exact transitions, or unrealizable reactions, so undecidability of the encoding need not govern deployed materials models.","failure_condition":"No enforceable fragment covers useful cases, UNKNOWN dominates decisions, or guarantee labels are routinely ignored downstream.","problem_falsifier":"The declared production class is proven finite or admits a checked total exact hazard-reachability algorithm under the actual representation and computation model.","intervention_falsifier":"In the shadow test, boundary mapping and routing do not reduce false SAFE labels or scope confusion relative to baseline, or they reject materially more valid cases without improving guarantee correctness.","risks":["A checked theorem may formalize the wrong chemical task.","Coarse over-approximation may flood users with false alarms.","Restricting expressiveness may omit rare but important chemistry.","Users may operationally interpret UNKNOWN as SAFE.","The impossibility hypothesis may be rhetorically promoted before proof."]},"null_rationale":null,"classification":{"candidate_kind":"TESTABLE_CONJECTURE","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":0.72,"generator_notes":"Closed-book structural transfer. The chemistry-specific undecidability premise is explicitly hypothetical pending a semantics-preserving reduction or constructive refutation."}