{"abstention_reason":null,"arm":"SUBSTRATE_DIVERSE_P2","candidate_id":"ensemble_and_population_level_equilibrium_versus_individual_level_heterogeneity__literature_literary_theory__SUBSTRATE_DIVERSE_P2","cell_id":"ensemble_and_population_level_equilibrium_versus_individual_level_heterogeneity__literature_literary_theory","decision":"PROPOSAL","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","proposal":{"actors":["Textual scholars and critical-edition editors","Rare-book conservators who authorize handling and illumination","Holding libraries that steward surviving copies","Bibliographers who distinguish impressions, gatherings, cancels, and press states","Readers whose interpretations depend on punctuation, wording, typography, or erased inscriptions"],"affected_objective":"Preserve a defensible edition-level base text while preventing its stable modal reading from being treated as the physically identical reading of every surviving copy.","arm":"SUBSTRATE_DIVERSE_P2","authority_safety":{"authorized_first_step":"A rare-book conservator may authorize non-contact, reversible multi-angle and narrowband imaging of preselected leaves in eight surviving copies from one verified impression; a textual scholar may compare masked captures without changing any catalog or edition record.","decision_authority":"The holding institution and conservator retain authority over object handling and illumination exposure. The designated textual editor retains authority over apparatus and interpretive claims after independent bibliographic review.","excluded_actions":["Destructive sampling of paper, ink, or binding material","Cleaning, erasing, humidifying, flattening, rebinding, or separating leaves","Using ultraviolet illumination without object-specific conservation approval and a recorded exposure limit","Automatically changing a base text, apparatus, catalog record, attribution, or provenance statement","Treating image enhancement, OCR, or automated segmentation as ground truth","Publishing ownership marks or reader annotations that create privacy, security, or provenance risks"],"halt_rollback":"Stop imaging if the cradle strains a binding, illumination produces detectable warming or conservation concern, registration requires physical pressure, or repeat captures are unstable. Return the volume unchanged, discard unvalidated classifications, retain the existing edition-level reading, and preserve only institution-approved raw captures and exposure records."},"baseline":"Editors commonly establish an edition-level reading from representative witnesses, existing collations, or reproductions and treat unreported copies of the same impression as materially equivalent. Local anomalies seen by unaided light may be dismissed as damage or transcription noise because no standardized physical measurement distinguishes weak ink, overprinting, stop-press correction, cancel insertion, abrasion, or erased inscription.","candidate_id":"ensemble_and_population_level_equilibrium_versus_individual_level_heterogeneity__literature_literary_theory__SUBSTRATE_DIVERSE_P2","causal_chain":["Define the ensemble as surviving copies bibliographically verified to belong to one impression, with individual copies as members and printed loci within specified gatherings as repeated measurement sites.","Define the macro equilibrium indicator as the proportion of sampled loci sharing the same modal mark geometry and reading across copies, considered stable when it remains inside a predeclared band across gatherings and leave-one-copy-out checks.","Place each volume in a conservation cradle and illuminate selected loci sequentially with calibrated visible narrowband, near-infrared, transmitted, and raking light within approved exposure limits.","Different inks, paper fibers, impressions, abrasions, overwriting, and inserted leaves interact differently with wavelength, transmission, and illumination angle, producing optical contrasts unavailable to ordinary frontal viewing.","Repeat captures and physical reference targets separate reproducible material signals from glare, page curvature, sensor noise, and registration artifacts.","Bibliographers classify only physically corroborated local states, such as weak original impression, later abrasion, stop-press correction, cancel leaf, overprint, or erased inscription, while retaining unresolved states as unresolved.","Each locus receives an equal-copy modal reading for the edition-level indicator, while copy-specific states and uncertainty remain attached to the individual physical witness; aggregation discards material sequence, minority states, and provenance-specific interventions.","The stable modal text is preserved when robustness checks pass, but a copy-level apparatus note is considered where a reproducible material state changes wording, punctuation, speaker attribution, typography, or the interpretation of reader intervention.","Later edition review monitors both modal textual concordance and the distribution of material witness states so additional copies can refine local notes without silently destabilizing the base text."],"cell_id":"ensemble_and_population_level_equilibrium_versus_individual_level_heterogeneity__literature_literary_theory","consequence":"A valid impression-level base text is overextended into a claim of copy-level identity, causing physically distinct readings or interventions to be erased and potentially attributing a printer's correction, material loss, or later reader's mark to the authorial text.","diversity_from_prior_proposals":"P1 concerns period-level narrative-mode generalization across novels and intervenes through passage annotation, equal-work aggregation, and close-reading review. This proposal concerns material non-identity among physical copies of one impression and intervenes through wavelength-, transmission-, and angle-dependent optical measurement. Its causal path runs from photon–material interaction to recovery of copy-specific marks and bounded textual claims, rather than from semantic passage classification to a corpus distribution.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","intervention":"Deploy a conservation-safe Multi-Angle Spectral Copy Comparator: a physical cradle, calibrated narrowband and near-infrared illumination, transmitted and raking-light geometry, reference targets, and a non-contact camera are used to measure the same textual loci across surviving copies of one impression. The resulting material contrasts preserve a stable modal base text while identifying physically corroborated copy-specific readings for bounded apparatus and interpretation.","mechanism_mapping":[{"counterfactual_removal":"Without wavelength-, transmission-, and angle-dependent illumination, weak, overwritten, abraded, or materially substituted marks that look alike under ordinary frontal light remain physically unresolved; the intervention loses its essential effect.","mechanism_slug":"multiangle_spectral_reflectance_measurement","role":"Create non-destructive optical contrast from differences in ink absorption, surface relief, paper transmission, and material layering."},{"counterfactual_removal":"Without the crosswalk, measured copy states remain isolated curiosities and cannot delimit what the stable impression-level reading does or does not license.","mechanism_slug":"micro_macro_crosswalk","role":"Connect equal-copy modal readings to the base text while retaining minority physical states, uncertainty, and information lost through aggregation."},{"counterfactual_removal":"Without stratified locus and copy selection, convenient clean leaves could establish macro stability while disputed gatherings, cancels, or edge cases remain unmeasured.","mechanism_slug":"stratified_sampling_review","role":"Sample control loci, disputed loci, multiple gatherings, and copies with different provenance or condition without treating those strata as textual causes."},{"counterfactual_removal":"Without copy-removal and gathering-level stress tests, one unusual witness or densely sampled section could appear to determine the supposedly stable modal text.","mechanism_slug":"equilibrium_stress_test","role":"Test whether the impression-level concordance remains stable under leave-one-copy-out checks and balanced sampling across gatherings."}],"nearest_rivals":["Traditional visual collation can identify conspicuous variants but cannot reliably discriminate faint ink, surface relief, transmission differences, or layered marks that require controlled optical interaction.","A conservation condition survey records damage and handling needs but does not relate a stable edition-level reading to interpretively consequential copy-level states.","Textual emendation changes or debates the base text; this proposal can preserve the base text while bounding claims about individual witnesses.","OCR or digital corpus collation compares encoded characters, whereas the essential evidence here is generated by direct measurement of ink, paper, relief, and layering on physical objects.","Variability characterization could inventory copy differences, but this proposal specifically tests them beneath an explicit, potentially valid impression-level textual equilibrium."],"negative_tests":{"intervention_falsifier":"The intervention is falsified for the pilot if blinded bibliographers using the calibrated captures do not distinguish predeclared material-state controls or adjudicate disputed loci more reproducibly than they do with standardized visible-light inspection, after excluding handling and exposure differences.","problem_falsifier":"The problem is falsified if existing copy-specific apparatus already bounds every decision-relevant claim, or if controlled measurements find no reproducible minority physical states that could change a reading or interpretation while the modal impression-level text remains stable.","risks":["Excess illumination or handling could damage fragile volumes.","Page curvature, show-through, glare, or sensor response could create false material distinctions.","Later damage may be mistaken for a production variant or authorially relevant state.","A small and institutionally clustered witness set may not represent all surviving copies.","Spectral enhancement may encourage false confidence where the physical signal remains ambiguous.","Ownership marks, erased annotations, or provenance evidence may be sensitive.","Editors may overvalue measurable material differences while neglecting linguistic and historical context."],"strongest_counterevidence":"If repeated multi-angle and spectral measurements show that apparent copy differences are solely later damage with no recoverable change to wording, punctuation, typography, or reader intervention, then the heterogeneity is physically real but not relevant to the literary decision."},"next_evidence_step":"With conservator approval, select eight copies from one verified impression, 200 randomly distributed control loci, and 20 loci already recorded as visually ambiguous. Capture each locus twice using calibrated frontal visible, two raking-light directions, transmitted light where binding permits, and approved narrowband near-infrared illumination. Give masked image sets and standardized visible-light photographs to two independent bibliographers, record state classifications and confidence, test repeatability and inter-reviewer agreement, calculate equal-copy modal concordance with leave-one-copy-out checks, and ask a textual editor whether any corroborated minority state changes a predeclared interpretive claim. Make no external record or edition changes during the pilot.","observable_state":"Across a bibliographically bounded impression, most sampled loci can sustain a stable equal-copy modal reading, while a smaller set may show repeatable copy-specific optical states clustered by gathering, leaf, or provenance. Controlled changes in wavelength, transmission, or illumination angle make those local states observable without implying that an individual copy shares every property of the modal text.","prior_art_status":"UNSEARCHED","problem":"A critical editor treats the stable modal text of a printed impression as though every surviving copy physically realizes that same reading. The aggregate base text may be valid, yet copy-specific press corrections, cancel leaves, weak impressions, abrasions, overprints, or erased annotations remain unresolved and are therefore either erased from interpretation or misattributed to the authorial text.","proposal_index":2,"remaining_contrastive_claim":"The proposal is warranted when an impression-level base text remains robust but controlled physical measurement reveals consequential minority witness states; unlike visual collation, condition surveying, OCR comparison, or base-text emendation alone, it preserves the macro reading while limiting claims about particular material copies.","revision_record":{"claim_changes":[],"conceptual_changes":[],"evidence_changes":[],"operational_changes":[],"parent_version":null,"progress_targets_addressed":["Created a problem independent of P1 at the level of physical textual witnesses rather than period-level narrative modes.","Made non-destructive optical instrumentation the essential causal substrate.","Specified a testable impression-level equilibrium, copy-level heterogeneity profile, and aggregation rule.","Bounded object-handling and editorial authority with reversible safeguards.","Added finite problem and intervention falsifiers and a conservation-safe evidence step."]},"schema_version":1,"structural_mapping":[{"archetype_element":"Ensemble Frame","domain_realization":"Surviving physical copies bibliographically verified as members of one printed impression; copies are ensemble members and selected textual loci are repeated measurement sites."},{"archetype_element":"Macro Equilibrium Indicator","domain_realization":"Equal-copy modal textual concordance across sampled loci and gatherings, with a predeclared stability band and leave-one-copy-out robustness check."},{"archetype_element":"Microstate Variability Profile","domain_realization":"Copy-specific spectral, transmission, and surface-relief states at each locus, including minority readings, material interventions, unresolved states, and measurement uncertainty."},{"archetype_element":"Aggregation Translation Rule","domain_realization":"Each copy contributes one classified reading per locus; the modal reading supports the impression-level text, while aggregation loses minority states, material sequence, provenance, and the distinction between production and later intervention."},{"archetype_element":"Level-of-Analysis Boundary","domain_realization":"The modal reading licenses a claim about the bounded impression, not a claim that every copy, leaf, or mark is identical; copy-level claims require direct physical evidence from that witness."},{"archetype_element":"Heterogeneity Relevance Test","domain_realization":"A minority state matters only when it is repeatable across permitted measurement geometries, survives artifact review, and changes wording, punctuation, typography, speaker attribution, provenance-sensitive interpretation, or the status of a reader intervention."},{"archetype_element":"Subgroup and Locality Map","domain_realization":"Material states are mapped by copy, gathering, leaf, forme position, and known provenance without inferring unrecorded ownership identities."},{"archetype_element":"Multi-Level Feedback Design","domain_realization":"The critical apparatus retains the stable impression-level reading while linking corroborated minority states to bounded copy-level notes and rechecking both levels when new witnesses are measured."},{"archetype_element":"Representative Case Guardrail","domain_realization":"Random control loci and predeclared ambiguous loci are sampled across gatherings and copies; neither a pristine exemplar nor a dramatic damaged leaf may stand in for the entire impression."}],"substrate_contract":{"counterfactual_independence":"Remove image databases, OCR, enhancement algorithms, dashboards, automated registration, and editorial workflow rules: the calibrated illumination, transmitted/raking geometry, camera, reference target, and direct comparison of raw captures still produce the material contrasts that reveal copy-specific states. Digital tools can organize or magnify the evidence but do not create its essential optical effect.","forbidden_channel_audit":"Governance limits exposure and reserves editorial authority; organizational process schedules access; software may register captures or tabulate classifications. None supplies the decisive contrast, which arises from physical interaction among illumination wavelength and angle, ink absorption, paper transmission, surface relief, and the camera sensor. No training, incentive, database, model, algorithm, information-routing system, or digital control is treated as the intervention's primary cause.","primary_allowed_process":"MEASUREMENT_INSTRUMENTATION"},"title":"Multi-Angle Spectral Copy Comparator","version":0},"schema_version":1}