{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp05_complete_proposal_portfolio20_20260803","cell_id":"negative_space_design__rhetoric","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_id":"negative_space_design__rhetoric__evidence_moats","proposal_index":5,"version":0,"title":"Evidence Moats for Multi-Claim Advocacy Sheets","problem":"A compact advocacy sheet can place several claims, statistics, qualifications, quotations, and citations close together without clearly indicating which evidence supports which assertion. Readers may group adjacent material by accidental proximity, assigning a source or qualifier to the wrong claim even though every necessary element is technically present.","actors":["Advocacy writer defining the claims and intended support relationships","Subject-matter reviewer checking evidence and qualifications","Visual or document designer implementing the sheet","Audience member evaluating the argument","Accessibility reviewer verifying equivalent nonvisual structure"],"observable_state":"A page contains multiple claims and evidence fragments separated by inconsistent gaps, while citations accumulate in a shared footer or edge column. In reader checks, people connect statistics, quotations, qualifications, or source numbers to neighboring claims differently from the writer's intended mapping.","consequence":"The sheet's visual rhetoric can imply evidentiary relationships that the underlying argument does not contain. Readers may overestimate support for one claim, miss a limitation attached to another, or be unable to reconstruct which source bears which argumentative load.","affected_objective":"Make the intended claim-evidence-qualification relationships in a multi-claim persuasive document perceptible without adding decorative marks or removing necessary evidentiary context.","intervention":"Adopt an Evidence Moat layout system. First partition the argument into explicit claim units, each containing one focal claim and its directly associated warrant or explanation, evidence, qualification, and citation. Keep those elements close using a defined within-unit spacing token. Surround every unit with a larger, unmarked blank band that no quotation, number, citation marker, ornament, or cross-unit connector may enter. The between-unit gap must remain visibly larger than every within-unit gap across print, presentation, and responsive layouts; if the surface cannot preserve that distinction at readable type sizes, continue the units on another page or switch to a labeled comparison table. Shared definitions and scope conditions remain in a separate context header. Digital and accessible versions encode the same units with headings, lists, and programmatic groups so evidentiary boundaries do not depend on sight alone.","structural_mapping":[{"archetype_element":"Attention Competition Map","domain_realization":"Claims, statistics, quotations, qualifications, footnotes, decorative callouts, and shared citations are mapped according to which elements compete or become accidentally associated when placed near one another."},{"archetype_element":"Omission Candidate","domain_realization":"Decorative material, duplicated slogans, cross-unit connectors, and shared footnote placement are removed or relocated when they intrude on the blank bands; claim-specific evidence and qualifications remain."},{"archetype_element":"Protected Empty Space","domain_realization":"A defined blank band surrounds each complete claim unit and is protected from later copy additions, citation overflow, branding, and ornamental fill."},{"archetype_element":"Positive Form Relationship","domain_realization":"The empty band makes the positive elements inside each unit read as a related claim-evidence-qualification bundle while separating them from neighboring bundles."},{"archetype_element":"Absence Boundary","domain_realization":"Reusable within-unit and between-unit spacing tokens specify where a moat begins, how wide it must remain, and what content may not cross it."},{"archetype_element":"Meaning-of-Absence Check","domain_realization":"The document system defines the larger gap as an argumentative boundary meaning that material on opposite sides has no implied support relationship; nonvisual grouping conveys the same meaning accessibly."},{"archetype_element":"Context Preservation Frame","domain_realization":"Shared definitions, decision scope, dates, authorship, and general methodological limitations remain in a clearly identified context header rather than being deleted to create space."},{"archetype_element":"Accessibility and Recoverability Guardrail","domain_realization":"Readable type, logical headings, programmatic grouping, full citations, and an alternate labeled table are preserved; overflow moves to another surface instead of being hidden or compressed."},{"archetype_element":"Clarity or Effect Test","domain_realization":"Readers must match each claim to its intended evidence, qualifier, and citation; the layout succeeds only if visual and nonvisual users recover the authored argumentative groupings."}],"mechanism_mapping":[{"mechanism_slug":"margin_and_gutter_system","role":"Codifies repeated within-unit gaps and larger protected inter-unit gutters so the separation survives different authors, formats, and content revisions.","counterfactual_removal":"Without a reusable spacing grammar, gaps are tuned inconsistently and later additions can consume the boundaries, causing identical distances to signal both association and separation."},{"mechanism_slug":"whitespace","role":"Uses proximity within a unit and blank separation between units as the direct perceptual signal for which claim, evidence, qualification, and citation belong together.","counterfactual_removal":"Without whitespace as the grouping signal, adjacent fragments can form accidental associations unless additional borders, colors, or labels are introduced."},{"mechanism_slug":"sparse_layout","role":"Limits each surface to the number of complete claim units that can fit while preserving readable text and moat dimensions, moving excess units rather than crowding them into the remaining area.","counterfactual_removal":"Without a limit based on perceptual capacity, content pressure can shrink the moats and type until the intended evidence groupings are no longer legible."}],"causal_chain":["The writer specifies the intended support and qualification relationship for every claim before layout begins.","The designer places each claim and only its associated evidence, qualification, and citation into one close-proximity unit.","A protected blank moat is placed between that unit and every neighboring argumentative unit.","Because the inter-unit gap is consistently larger than gaps inside a unit, proximity signals association and separation signals the absence of an evidentiary relationship.","Content that cannot fit without breaching a moat moves to another surface or an explicit comparison structure rather than consuming the boundary.","Accessible structure reproduces the same group boundaries for readers who cannot perceive visual whitespace.","Reader matching tests determine whether the authored argumentative relationships, rather than accidental adjacency, govern interpretation."],"baseline":"A dense one-page fact sheet in which several claims, statistics, quotations, and qualifications share columns, sidebars, or a consolidated citation footer. Spacing is adjusted locally to make all content fit rather than used consistently to encode evidentiary relationships.","nearest_rivals":["Borders or shaded cards around each claim, which make grouping explicit by adding visible marks rather than using protected absence.","Numbered citation markers and a shared reference list, which identify sources but may not clarify which qualification, statistic, or warrant belongs to each claim.","Color coding, which can distinguish bundles visually but introduces accessibility, reproduction, and semantic-consistency requirements.","A one-claim-per-page format, which strongly isolates claims but may prevent necessary comparison and expand the document substantially.","A labeled evidence matrix, which makes relationships explicit and may be preferable for dense expert comparison but changes the rhetorical sheet into a tabular analytic artifact."],"remaining_contrastive_claim":"The proposal's testable distinction is a reusable layout grammar in which protected, unmarked separation carries the claim-boundary signal: content within a moat is asserted as one argumentative bundle, while content across a moat has no implied evidentiary relationship.","authority_safety":{"decision_authority":"The document owner may adopt the layout and move material across pages, but the subject-matter owner controls claim-support mappings and qualifications. Accessibility reviewers may require explicit labels or an alternate representation when whitespace alone is insufficient.","authorized_first_step":"Reformat one fictional or non-live six-claim advocacy sheet into an Evidence Moat version while preserving every claim, qualification, source, and shared contextual statement from the baseline.","excluded_actions":["Deleting contrary evidence, uncertainty, limitations, or required disclosures to create blank space","Separating a qualification from the claim it limits","Using isolation to imply that weak evidence is more authoritative","Shrinking type or interactive targets to preserve moat dimensions","Relying on visual spacing as the only structural signal in accessible documents","Moving citations so far from claims that their relationship becomes harder to verify","Applying the system where dense side-by-side comparison is the primary task and a labeled table is more interpretable"],"halt_rollback":"Stop the layout pilot if any required evidence or qualification cannot remain with its claim, readable sizing cannot be maintained, or accessibility testing cannot recover the same group structure. Restore the complete baseline content and switch to explicitly bordered, labeled, or tabular units while documenting which moat constraint failed."},"negative_tests":{"strongest_counterevidence":"Explicit borders, headings, or tables may communicate claim boundaries more reliably than unmarked spacing, particularly on small screens, in low-vision use, or when readers must compare dense evidence across claims.","problem_falsifier":"Reader checks show that claim-evidence and claim-qualification mappings are already recovered consistently, or analysis shows that misunderstanding results from ambiguous reasoning or weak evidence rather than spatial crowding.","intervention_falsifier":"Readers using the Evidence Moat version continue to assign evidence or qualifications to unintended claims, interpret isolated units as unrelated to shared context, or require more navigation without a clearer reconstruction of the argument.","risks":["Large gaps may fragment an argument whose claims genuinely depend on one another.","Isolated units may make evidence appear more independent than it is.","Moving overflow to later pages may conceal material from readers who stop early.","Repeated spacing may fail when translated text expands or responsive screens become narrow.","A sparse sheet may appear unfinished or may exaggerate the salience of the remaining claims.","Programmatic group labels may disclose relationships differently from the visual version.","Designers may protect visual regularity at the expense of qualification length or citation completeness."]},"next_evidence_step":"Prepare baseline and Evidence Moat versions of the same fictional six-claim sheet, preserving identical wording, evidence, qualifications, citations, typography size, and context. Randomly assign 12 to 16 consenting readers to one version and ask them to match each claim with its supporting evidence, limiting qualification, and source before collecting confidence and perceived grouping explanations. Conduct a separate screen-reader walkthrough of the programmatic units and test the narrowest intended responsive layout. Use the results only to assess grouping interpretation, accessibility, and whether an explicit bordered or tabular fallback is required.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Proposal 1, Protected Warrant Silence for Live Argument Testing, withholds moderator speech to reveal whether participants supply an unstated premise; Evidence Moats instead organize already-stated claims and evidence in a finished artifact and do not elicit inference through conversational silence. Proposal 2, Reserved Residual Field for Binary Deliberation, protects one participant-fillable region to expose alternatives outside a supplied option partition; Evidence Moats use repeated unfillable gaps to prevent false evidentiary association, not to expand an option set. Proposal 3, Correction Landing Window for Institutional Channels, creates a temporal pause in future publication after a correction; Evidence Moats alter simultaneous spatial grouping inside one document and do not govern message cadence. Proposal 4, Unannotated First Pass for Rhetorical Criticism, temporarily hides recoverable commentary so readers encounter a primary source before interpretation is supplied; Evidence Moats hide no substantive interpretive layer and instead keep each claim's evidence, qualifier, and citation visible together. This proposal is independently adoptable as a visual and accessible document-layout system.","revision_record":{"parent_version":null,"progress_targets_addressed":[],"conceptual_changes":[],"operational_changes":[],"evidence_changes":[],"claim_changes":[]}}