{"actors":["Literary translator responsible for rendering the work","Translation editor responsible for cross-passage consistency and source-text fidelity review","Publisher production lead who controls copyedit and typesetting lock","Author or rights holder when contractual approval applies"],"affected_objective":"Preserve source-text distinctions, recurrent motifs, voice, and formal effects across a translated work without forcing expensive global revision after translation rules and pages have been locked.","arm":"ORDINARY_DIVERSE_P2","authority_safety":{"authorized_first_step":"The translator may compare candidate rendering rules on duplicate worksheets made from a public-domain source and nonproduction draft translations; no source text, production manuscript, contractually approved wording, or live style sheet is changed.","decision_authority":"The accountable translator chooses the rendering policy, the translation editor approves cross-work consistency, and the production lead controls manuscript lock; any contractually reserved author or rights-holder approval remains mandatory. Forecasts are advisory and cannot resolve contested interpretation automatically.","excluded_actions":["Automatically replacing the translator's wording in the production manuscript","Treating the forecasted constraint score as proof of a uniquely correct translation","Editing or normalizing the source text to remove translation difficulty","Training an external system on confidential manuscripts without authorization","Overriding author, rights-holder, or translator approvals established by contract","Concealing uncertainty or unresolved interpretive alternatives from accountable reviewers"],"halt_rollback":"Halt if the preview omits source occurrences, invents textual correspondences, produces unstable classifications under minor annotation changes, or exposes confidential text. Discard worksheet-only recommendations, retain the existing manuscript and style sheet, and return disputed choices to human review."},"baseline":"The translator establishes recurring terms, voice conventions, and formal strategies while working sequentially, records some decisions in a glossary or style sheet, and reconciles conflicts when later passages or editorial review reveal them.","candidate_id":"predictive_precommitment_correction__literature_literary_theory__ORDINARY_DIVERSE_P2","causal_chain":["The translator identifies a proposed rendering rule—such as a recurring lexeme, pronoun distinction, register marker, line-length pattern, or transliteration convention—before the affected chapters are locked.","A source-text occurrence map locates later passages in which that rule will interact with motif recurrence, polysemy, rhyme, meter, character voice, or plot-dependent ambiguity.","A constrained preview applies each candidate rule to representative future occurrences and predicts semantic collapses, formal conflicts, voice drift, and required exceptions, with uncertainty recorded for disputed source annotations.","The translator compares each candidate's predicted cross-work consequence with a predefined tolerance envelope for preserved distinctions, recurrence, formal compatibility, voice consistency, and exception burden.","If the predicted gap crosses a threshold, the translator adjusts the rendering, scope, exception rule, rhythmic template, sequencing of passage work, or timing of commitment.","A chapter-lock gate requires the translator and translation editor to accept, revise, delay, escalate, or reject the proposed rule, while permitting a documented human override for interpretive reasons.","As the remaining work is translated and edited, reviewers record which predicted conflicts occurred, which did not, and which unpredicted conflicts required global revision.","Those forecast errors recalibrate occurrence classifications, thresholds, and confidence or demote the preview method to a nonbinding concordance aid."],"cell_id":"predictive_precommitment_correction__literature_literary_theory","consequence":"After a rendering rule has propagated through the manuscript and copyedit or typesetting has begun, discovering that it erases a later distinction, breaks a recurrent formal pattern, or destabilizes a character's voice can require coordinated revision across many passages, notes, indexes, and page proofs.","diversity_from_prior_proposals":"This independent opportunity addresses cross-work constraint propagation in literary translation. Its intervention previews how a candidate translation rule behaves in later source passages and pre-corrects the rule before manuscript lock; it does not configure critical-edition paratext or predict readers' interpretations.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","intervention":"Insert a whole-work translation-rule preflight before chapter lock. For each consequential recurring choice, construct a source-text occurrence map, apply candidate rendering rules to a bounded set of later contexts, forecast semantic, formal, and voice conflicts with explicit uncertainty, and compare the forecast with a declared tolerance envelope. Require a recorded accept, revise, delay, escalate, or reject decision; permit only translator-approved changes to renderings, rule scope, exceptions, rhythmic templates, work sequence, or lock timing; then compare forecasted conflicts with those encountered during subsequent translation and editing to recalibrate the preview.","mechanism_mapping":[{"counterfactual_removal":"Without projecting each rendering rule into later source contexts, the translator cannot inspect its likely global consequences until the rule has already propagated through the manuscript.","mechanism_slug":"precommitment_what_if_simulation","role":"A source-occurrence map and candidate renderings create alternative whole-work previews of how local translation choices interact with later semantic and formal constraints."},{"counterfactual_removal":"Without a lock decision tied to the forecast, identified cross-work conflicts can remain informational while the affected chapters proceed into copyedit or typesetting.","mechanism_slug":"staged_commitment_gate","role":"Chapter or style-sheet commitment is conditioned on a recorded accept, revise, delay, escalate, or reject decision with an accountable override."},{"counterfactual_removal":"Without comparing predicted conflicts with later adjudicated translation problems, missed dependencies and false alarms cannot recalibrate the occurrence map, thresholds, or confidence.","mechanism_slug":"forecast_error_backtest","role":"Subsequent translation and editorial decisions supply a trace of realized conflicts against which the preflight's predictions are scored and revised."}],"nearest_rivals":["A translation glossary records chosen equivalents, whereas this proposal additionally projects a proposed rule through future contexts, evaluates predicted conflicts against a tolerance envelope, and gates commitment.","A concordance locates repeated source forms, whereas the proposed preview represents interactions among recurrence, polysemy, voice, plot context, and formal constraints and converts the predicted gap into an adjustable translation decision.","Back-translation can expose divergence in completed target text, whereas this intervention evaluates alternative rendering policies before they propagate through the manuscript.","Ordinary translation editing can reconcile inconsistencies during review, whereas the proposed causal structure places consequence modeling and rule adjustment before chapter lock and preserves later review outcomes for calibration."],"negative_tests":{"intervention_falsifier":"In a blinded crossover on comparable passage sets, preflight-selected rendering rules do not produce fewer adjudicated global rule reversals or fewer missed source distinctions than translator decisions made with the baseline glossary and source text, or selections reverse under immaterial annotation changes.","problem_falsifier":"Candidate rendering policies produce no distinguishable downstream semantic, formal, or voice conflicts across the mapped occurrences, or all affected passages remain cheaply revisable together until final release, so no meaningful costly commitment point exists.","risks":["The occurrence map may mistake surface repetition for meaningful recurrence or miss relations that depend on interpretation.","A numeric tolerance envelope may flatten legitimate ambiguity and encourage mechanical lexical consistency.","Previewing later passages may bias the translator toward retrospective coherence unsupported by the source.","Optimizing recurring terms can damage local rhythm, idiom, humor, or characterization.","Editors may game the exception threshold to secure production lock.","Forecast confidence may displace accountable interpretive judgment.","Confidential or embargoed source and translation text could be exposed through an unauthorized modeling tool."],"strongest_counterevidence":"Literary translation choices may be too context-sensitive and mutually constitutive for isolated rule previews to predict their eventual consequences; the act of translating intervening passages can change the translator's understanding so substantially that an early global model performs worse than revisable sequential judgment."},"next_evidence_step":"Select one public-domain short work containing one recurrent polysemous term and one recurrent formal feature. Have a translator define two candidate policies for each feature, map all occurrences, and—using only the first half plus source-level annotations—predict conflicts in the second half before translating it. Translate the second half under blinded policy labels, then have two independent reviewers adjudicate preserved distinctions, formal conflicts, voice disruptions, and required global revisions using a rubric fixed in advance. Compare predictions with adjudications and the translator's unaided baseline choice; make no production-manuscript changes.","observable_state":"Before lock, the accountable actors can observe the proposed rendering rule; every mapped source occurrence; contextual annotations; candidate target renderings; predicted semantic collapses, motif breaks, formal conflicts, voice shifts, and exceptions; uncertainty and annotation sensitivity; tolerance thresholds; recommended adjustments; and recorded overrides. During later translation and editing, they can observe adjudicated conflicts, unpredicted dependencies, false alarms, global reversals, and forecast error.","prior_art_status":"UNSEARCHED","problem":"A literary translator must often choose recurring lexical, grammatical, vocal, and formal conventions before all of their downstream interactions have been worked through. A locally defensible rendering can later collapse a source distinction, disrupt a motif, make a rhyme or meter constraint unsatisfiable, or create inconsistent character voice. The rule remains adjustable before chapter and style-sheet lock, but later correction can propagate across the manuscript and production apparatus; no structured gate currently connects a preview of those cross-work consequences to the initial translation decision in this proposal.","proposal_index":2,"remaining_contrastive_claim":"The proposal's contrastive claim is structural rather than novel: a candidate translation rule is projected through future source contexts, its predicted global consequence and uncertainty are compared with an explicit tolerance envelope, the rule is adjusted before manuscript lock, and later adjudicated conflicts recalibrate the forecast. A glossary, concordance, completed-text review, or ungated preview lacks at least one of these links.","revision_record":{"claim_changes":["Initial version; no claims of novelty, prevalence, demand, or effect size are made."],"conceptual_changes":["Initial version maps predictive precommitment correction to prospective management of cross-work constraints in literary translation."],"evidence_changes":["No external evidence was consulted; prior art remains unsearched."],"operational_changes":["Initial test is bounded to one public-domain short work, two recurrent features, two candidate policies per feature, duplicate worksheets, and no production changes."],"parent_version":null,"progress_targets_addressed":["Independent problem, intervention, and causal path distinct from paratext design","Explicit translation commitment point and adjustable variables","Future-context consequence preview with uncertainty","Accountable manuscript-lock gate and human override","Realized-conflict calibration trace","Falsifiers, risks, safeguards, and a bounded first test"]},"schema_version":1,"structural_mapping":[{"archetype_element":"Commitment point after which correction becomes expensive or disruptive","domain_realization":"Chapter and translation-style-sheet lock before copyedit or typesetting, after which a changed rendering rule can require coordinated revisions across passages and production artifacts."},{"archetype_element":"Intended action, target state, and tolerance envelope","domain_realization":"Adopt a specified recurring translation rule while keeping preserved source distinctions, motif recurrence, formal compatibility, character voice, and exception burden within declared bounds."},{"archetype_element":"Predictive consequence model upstream of commitment","domain_realization":"An occurrence map and constrained rendering preview estimate how alternative rules will behave in later semantic, narrative, vocal, and prosodic contexts."},{"archetype_element":"Context state input","domain_realization":"Source wording, speakers, narrative position, plot-dependent information, lexical relations, motif links, rhyme or meter positions, and uncertainty in interpretive annotation."},{"archetype_element":"Predicted gap signal with uncertainty","domain_realization":"Forecasted loss of a distinction, broken recurrence, formal incompatibility, voice drift, or excessive exception count relative to the tolerance envelope, accompanied by confidence and annotation sensitivity."},{"archetype_element":"Adjustable control variables","domain_realization":"Target lexeme, syntax, pronoun or register marker, transliteration, rhythmic template, rule scope, exception list, order of passage work, and timing of chapter lock."},{"archetype_element":"Pre-correction rule","domain_realization":"When a forecasted violation crosses a predefined threshold, revise or narrow the rendering policy, add an explicit exception, translate a diagnostic passage earlier, or delay commitment; under substantial uncertainty, prefer reversible trials."},{"archetype_element":"Staged commitment gate and override","domain_realization":"The translator and translation editor record acceptance, revision, delay, escalation, rejection, or a reasoned human override before the rule and affected chapters are locked."},{"archetype_element":"Post-action feedback and recalibration","domain_realization":"Conflicts adjudicated during subsequent translation and editing are compared with predicted conflicts to update mappings and thresholds or reduce the preview's authority."},{"archetype_element":"Safeguards against drift, false confidence, and gaming","domain_realization":"Complete occurrence logs, visible uncertainty, sensitivity checks, independent review, fixed thresholds, recorded exceptions and overrides, confidentiality controls, and retained translator authority."}],"title":"Whole-Work Constraint Preflight for Literary Translation Rules","version":0}