{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"predictive_residual_processing__religious_studies_theology:P2:v0","cell_id":"predictive_residual_processing__religious_studies_theology","search_queries":["theological commission doctrinal document drafting revision review guidelines consultation official","computational theology ontology doctrinal consistency argument mapping research","church doctrinal document revision unintended consequences review consultation theology","argument mapping theology doctrinal statements software ontology","site:usccb.org doctrinally sound catechetical materials protocol review conformity checklist","theology ontology doctrinal statements knowledge graph semantic web peer reviewed","computational theology logical consistency doctrinal claims automated reasoning paper","\"doctrinal\" \"argument map\" theology","Argument Interchange Format ontology standard official specification AIF ontology","computational hermeneutics ontology argument formalization theology Benzmuller paper pdf","EUR-Lex GDPR Article 9 religious beliefs special categories official","NIST AI RMF human oversight documentation testing official"],"sources":[{"source_id":"S1","title":"Catechetical Accompaniment Process: Assessing Materials for Conformity to the Catechism of the Catholic Church","publisher":"United States Conference of Catholic Bishops","url":"https://www.usccb.org/committees/catechism/catechetical-accompaniment-process","source_class":"OFFICIAL_GUIDANCE","publication_date":"1997-09","accessed_at":"2026-08-03","claims_supported":["An authorized theological-review body uses a standard instrument to assess new texts against a designated prior commitment source.","The process requires authenticity, completeness, wider doctrinal context, and relationships among teachings rather than contradiction checking alone.","The protocol was revised after consultation with publishers, bishop reviewers, and consultants."]},{"source_id":"S2","title":"Archbishop Daniel Buechlein Report, June 1997","publisher":"United States Conference of Catholic Bishops","url":"https://www.usccb.org/beliefs-and-teachings/what-we-believe/catechism/archbishop-daniel-buechlein-report-june-1997","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"1997-06","accessed_at":"2026-08-03","claims_supported":["A single catechetical-series review required approximately 400 hours across committee members, a bishop-chair, expert reviewers, and staff.","The committee reported increased review demand and recruited additional review-team chairs.","Reviewed materials displayed recurring doctrinal incompleteness, imprecision, and weak representation of relationships among commitments."]},{"source_id":"S3","title":"The Role and Functions of Doctrinal Commissions","publisher":"Dicastery for the Doctrine of the Faith, Holy See","url":"https://www.vatican.va/roman_curia/congregations/cfaith/ladaria-ferrer/documents/rc_con_cfaith_doc_20190115_commissioni-dottrinali_en.html","source_class":"OFFICIAL_GUIDANCE","publication_date":"2019-01-15","accessed_at":"2026-08-03","claims_supported":["Episcopal doctrinal commissions are identifiable authorized adopters for doctrinal-review support.","A commission is accountable to its episcopal conference, is consultative, and cannot issue public statements without explicit authorization.","Theologians may provide expertise while bishops retain doctrinal authority.","Official guidance describes a need and benefit for commissions that evaluate doctrinal issues and publications while respecting prior decisions."]},{"source_id":"S4","title":"A Computational-Hermeneutic Approach for Conceptual Explicitation","publisher":"arXiv; authors affiliated with Freie Universität Berlin and University of Luxembourg","url":"https://arxiv.org/abs/1906.06582","source_class":"PRIMARY_RESEARCH","publication_date":"2019-06-15","accessed_at":"2026-08-03","claims_supported":["Computer-supported hermeneutics can formalize natural-language arguments and expose tacit conceptual commitments.","The demonstrated approach operates iteratively over networks of supporting and attacking arguments.","The work supports technical plausibility but acknowledges challenges in full automation and does not establish doctrinal-drafting workflow benefits."]},{"source_id":"S5","title":"The Argument Interchange Format (AIF) Specification","publisher":"Argumentation Research Group, University of Dundee","url":"https://www.arg-tech.org/wp-content/uploads/2011/09/aif-spec.pdf","source_class":"STANDARD","publication_date":"2011-09","accessed_at":"2026-08-03","claims_supported":["A mature argument-graph specification represents propositions, inference, conflict, preference, premises, assumptions, and exceptions.","AIF provides OWL and SQL representations and supports exchange among argument tools.","Graph representation and provenance attributes are established prior art for the proposed commitment graph, but AIF does not provide predictive residual prioritization or fallback review."]},{"source_id":"S6","title":"DoctrineLab — Comparative Theology Research Platform","publisher":"DoctrineLab","url":"https://doctrinelab.org/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"undated","accessed_at":"2026-08-03","claims_supported":["A first-party theology platform advertises doctrine dependency maps, evidence layers, source transparency, and AI-assisted comparative reading.","The product is adjacent prior art for structured doctrinal comparison and downstream-dependency exploration.","The public description does not establish pre-review prediction, residual reconstruction, blinded raw-text audits, or measured attention savings in an authorized drafting process."]},{"source_id":"S7","title":"Regulation (EU) 2016/679, Article 9: Processing of Special Categories of Personal Data","publisher":"EUR-Lex, European Union","url":"https://eur-lex.europa.eu/legal-content/DE-EN-SL/ALL/?from=EN&uri=CELEX%3A32016R0679","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2016-04-27","accessed_at":"2026-08-03","claims_supported":["Personal data revealing religious or philosophical beliefs are special-category data under GDPR Article 9.","A religious nonprofit may process qualifying member-related data in legitimate activities only with safeguards and limits on external disclosure, subject to applicable conditions.","Archive access, reviewer-attribution, affected-community information, and cross-border deployment require jurisdiction-specific legal review."]},{"source_id":"S8","title":"Handbook on the Conformity Review Process","publisher":"United States Conference of Catholic Bishops","url":"https://www.usccb.org/resources/CR-Handbook-only-FINAL.pdf","source_class":"OFFICIAL_GUIDANCE","publication_date":"2012","accessed_at":"2026-08-03","claims_supported":["An established conformity review may take six to twelve months for initial review.","The workflow uses publisher self-review, a bishop-chair, theologian or catechist reviewers, staff support, formal reports, and multiple authority levels.","The handbook identifies incomplete or inaccurate doctrinal content as the reason for review and describes comparison against an authoritative source as the mechanism for finding deficiencies.","The process provides an operational comparator and demonstrates that institutional authority and complete-text review cannot be delegated to software."]}],"problem_evidence":{"support":"MODERATE","rationale":"Official USCCB evidence shows labor-intensive, months-long theological conformity review, increased demand, and recurring incompleteness or imprecision. It therefore supports a consequential review problem and scarce expert attention. However, the evidence concerns catechetical conformity review, not revision of confessions or liturgical standards, and it does not directly measure how much effort is wasted on predictable restatement or how often redlines miss downstream semantic consequences.","source_ids":["S1","S2","S8"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Doctrinal commissions, episcopal conferences, bishops, and expert reviewers are identifiable adopters and authorizers with explicit mandates to examine doctrinal questions and publications. USCCB has maintained structured review processes and reports publisher cooperation. No source expresses demand for predictive-residual presentation specifically, and adoption would require approval from each institution's governing authority.","source_ids":["S1","S3","S8"]},"prior_art":{"proximity":"ADJACENT_PRIOR_ART","closest_analogues":[{"name":"USCCB conformity-review protocol","similarity":"Uses an authoritative doctrinal baseline, structured criteria, expert review, reports, correction cycles, and episcopal authorization to detect incompleteness and inconsistency.","remaining_difference":"It performs comprehensive criterion-based review rather than held-out prediction, residual reconstruction, consequence-weighted prioritization, independent residual-blind audits, and automatic fallback.","source_ids":["S1","S2","S8"]},{"name":"Computational hermeneutics","similarity":"Formalizes natural-language argumentative discourse, makes tacit commitments explicit, and reasons over networks of supporting and attacking arguments with humans in the loop.","remaining_difference":"It does not target an institution's draft revision workflow, compress review into expected profiles plus residuals, preserve plural institutional models, or test deliberative-time savings against full review.","source_ids":["S4"]},{"name":"Argument Interchange Format","similarity":"Provides a reusable graph ontology for propositions, inference, conflict, preference, assumptions, exceptions, and provenance-like attributes.","remaining_difference":"It is a representation and interchange specification, not a predictive review protocol with model versions, residual thresholds, held-out validation, raw-text auditing, or authority-sensitive fallback.","source_ids":["S5"]},{"name":"DoctrineLab","similarity":"Offers doctrinal dependency maps, layered evidence, source transparency, and AI-assisted comparison across theological materials.","remaining_difference":"The public product description does not show authorized document drafting, temporally frozen institutional baselines, typed prediction residuals, matched-reviewer evaluation, mandatory bypasses, or measured reconstruction fidelity.","source_ids":["S6"]}],"distinctive_claim_remaining":"Within one authorized revision episode, a temporally frozen and pluralized commitment graph can predict enough of held-out draft sections that expectation-plus-typed-residual review reduces total expert time versus complete-text review plus redlining, while remaining non-inferior for reconstruction of the complete descriptive commitment profile and producing zero misses in predeclared mandatory-review passages. The coordinated claim—not graphing, conformity review, or automated reasoning individually—remains unverified.","confidence":"MODERATE"},"implementation_evidence":{"support":"MODERATE","rationale":"Argument-graph standards, computational-hermeneutic research, and a first-party doctrinal-mapping product make a read-only prototype technically plausible. Existing church workflows establish expert roles, reports, source baselines, confidentiality, and retained human authority. The difficult components are empirical and socio-technical: reliable pre-review semantic prediction, plural interpretive mappings, inter-reviewer agreement, resistance to anchoring, secure access to restricted archives, and institution-specific privacy and authority approval.","source_ids":["S3","S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"The documented comparator can consume roughly 400 expert/staff hours and six to twelve months, so safe reductions in deliberative burden could be meaningful; the addressable setting is nevertheless specialized.","source_ids":["S2","S8"]},"stakeholder_pull":{"score":3,"rationale":"Authorized review bodies visibly exist and sustain structured workflows, but no adopter has requested predictive-residual tooling or committed resources to a pilot.","source_ids":["S1","S3","S8"]},"incremental_advantage":{"score":3,"rationale":"Residual prioritization could add attention savings and traceable consequence propagation beyond checklists and argument graphs, but neither time savings nor non-inferior issue recall has been demonstrated.","source_ids":["S4","S5","S6","S8"]},"distinctiveness_plausibility":{"score":3,"rationale":"Search found all major components separately but no direct source for their proposed combination in authorized doctrinal drafting. This is only a bounded-search contrast, not a world-novelty determination.","source_ids":["S1","S4","S5","S6"]},"technical_implementability":{"score":3,"rationale":"Graphs, formalized arguments, provenance, comparison, and read-only interfaces are implementable; dependable semantic prediction across contested theological interpretations remains uncertain.","source_ids":["S4","S5","S6"]},"adoption_authority_feasibility":{"score":3,"rationale":"The adopter and chain of authority are explicit, and the shadow test preserves episcopal decision authority. Approval, confidentiality, archive access, and local governance remain institution-specific barriers.","source_ids":["S3","S7","S8"]},"evidence_readiness":{"score":2,"rationale":"Public evidence establishes the neighboring workflow but not residual prevalence, predictive accuracy, anchoring effects, subgroup or interpretive-position recall, or net review-time benefit. Decisive evidence requires a proprietary archive and expert fieldwork.","source_ids":["S1","S2","S8"]},"safety_net_benefit":{"score":4,"rationale":"Complete-text access, residual-blind sampling, mandatory passage bypasses, version checks, and full-review fallback directly address model lock-in, although their effectiveness is untested.","source_ids":["S3","S7","S8"]},"scalability":{"score":2,"rationale":"The software pattern could be reused, but commitment graphs, interpretive plurality, authority rules, confidential sources, and validation thresholds must be rebuilt for each institution and document family.","source_ids":["S3","S5","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"One retrospective shadow study using 24 sections, two documented interpretive mappings, matched residual-led and complete-text reviewers, independent adjudication, privacy review, analysis, and a rejection-oriented report.","confidence":"MODERATE","assumptions":["The institution already possesses a usable completed revision archive and grants access without digitization or licensing expense.","Approximately 600-1,200 combined theologian, curator, adjudicator, analyst, and project-management hours are required.","No production integration or current doctrinal decision is included."],"source_ids":["S2","S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Build a secure read-only minimum viable workspace with graph authoring, provenance, versioning, residual dossiers, complete-text expansion, audit sampling, role controls, export, and fallback for one institution and document family.","confidence":"LOW","assumptions":["A small product and engineering team works for six to nine months with continuing theological curation.","Existing identity management and document repositories can be integrated rather than replaced.","No automated doctrinal decision-making or cross-institution model is built."],"source_ids":["S4","S5","S7","S8"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Governed launch for one institution covering security and privacy assessment, archive ingestion, two interpretive mappings, reviewer training, acceptance testing, support, and parallel full review for an initial revision cycle.","confidence":"LOW","assumptions":["The launch remains shadow or decision-support only until non-inferiority is demonstrated.","Formal theologian and authority time is included as resource-equivalent cost.","Restricted-content remediation or bespoke multilingual modeling could move the cost above this band."],"source_ids":["S3","S7","S8"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Annual model stewardship, source and permission maintenance, secure hosting, support, independent audits, expert review honoraria, drift checks, incident response, and one to three bounded revision episodes.","confidence":"LOW","assumptions":["One technical steward and one theological or archival steward are retained, with part-time security, legal, and expert-review support.","Every institution maintains its own baseline and authority process.","Full-review fallback capacity remains funded rather than counted as an exceptional external cost."],"source_ids":["S2","S3","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Official evidence documents lengthy, labor-intensive review and recurring doctrinal incompleteness or imprecision, although residual-specific prevalence is unmeasured.","source_ids":["S1","S2","S8"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"Episcopal conferences, doctrinal commissions, bishops, and established review subcommittees have explicit mandates and authority boundaries for this work.","source_ids":["S3","S8"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The candidate makes a contrastive non-inferiority-plus-time-saving claim against complete-text review and redlining, with zero tolerance for mandatory-passage misses.","source_ids":["S4","S5","S6","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A retrospective, temporally held-out, 24-section matched-reviewer study can test reconstruction, issue recall, time, interpretive disparity, and fallback without changing an adopted document.","source_ids":["S2","S4","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"UNCERTAIN","reason":"The design preserves human authority and full-review fallback, but archive permissions, confidentiality, special-category religious-belief data, affected-community representation, and anchoring risk require institution-specific review before testing.","source_ids":["S3","S7","S8"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The bands are broad and assumption-labeled, anchored by documented 400-hour and six-to-twelve-month review workflows; software and recurring estimates remain low-confidence.","source_ids":["S2","S8"]}},"next_evidence_step":"With one willing authorized institution, pre-register a retrospective shadow study on a completed revision archive. Fix a cutoff before the first draft; build the commitment graph only from earlier authorized sources; document two interpretive mappings; use 12 sections to define the relation vocabulary and freeze 12 untouched sections for testing. Randomize and counterbalance qualified reviewers between (A) residual-led review with required full-context expansion and (B) complete-text review plus redline and ordinary issues log. A separate blinded panel reviews a random sample and every rights-, discipline-, safeguarding-, membership-, pastoral-treatment-, or community-impact passage in full. Primary outcomes are total expert time and non-inferiority in adjudicated material-issue recall and full-profile reconstruction; report interpretive-position recall separately. Falsify the intervention upon any mandatory-passage miss, systematic miss for either interpretive mapping or affected community, reconstruction outside the predeclared tolerance, version/provenance failure, unauthorized disclosure, increased total effort after graphing and audits, or no meaningful time advantage. Because the archive and expert judgments are proprietary and performance must be observed, this step cannot be completed by further web research.","blocking_evidence":["No measured prevalence of predictable restatement versus genuinely novel content in completed doctrinal revisions.","No held-out estimate of commitment-profile prediction accuracy or reconstruction fidelity.","No evidence that residual-led reviewers retain non-inferior recall after accounting for anchoring and context-expansion behavior.","No subgroup analysis for documented interpretive positions or materially affected communities.","No institution-specific authorization, archive-access determination, privacy basis, retention policy, or confidentiality assessment.","No measured total cost including graph construction, disputes, audits, fallback, and model stewardship.","No evidence that mandatory bypasses and full-review fallback work reliably in practice.","No adopter commitment or funder expression specific to this intervention."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search supports only a contrast against the eight reviewed sources: established conformity review, argument-graph specifications, computational hermeneutics, and doctrinal dependency products are adjacent, while the full residual-review combination was not found. World novelty, patentability, freedom to operate, market size, and realized impact remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Secure written authorization and lawful access to one completed revision archive without expanding access to restricted records.","Pre-register material-issue definitions, reconstruction tolerance, mandatory-bypass classes, non-inferiority margin, time-saving threshold, and stopping rules.","Complete the 12-section held-out matched-reviewer comparison against full review plus redlining.","Demonstrate zero mandatory-passage misses and report recall separately for both interpretive mappings and affected-community concerns.","Measure total resource-equivalent effort including curation, graphing, dispute resolution, audits, context expansion, and fallback.","Obtain independent privacy, confidentiality, security, and authority approval before any operational pilot."],"reason":"Bounded web research has established a real neighboring review burden, credible authorities, and adjacent technical prior art, but it cannot resolve the decisive causal claim. The remaining evidence requires proprietary archival material and observed expert-review performance, so further desk research would not verify safety, non-inferiority, or net attention savings."},"proposal_index":2}