{"schema_version":1,"research_id":"eoa_inverse_innovation_exp05_external_evaluation_20260803","source_assessment_id":"negative_space_design__linguistics_semiotics:P3:v0","cell_id":"negative_space_design__linguistics_semiotics","search_queries":["linguistic landscape survey methodology absence signs sampling locations completeness","linguistic landscape data collection geotagged signs survey methodology fieldwork","linguistic landscape corpus database map signs research project official","linguistic landscape methodology absence missing data public signs","site:dwc.tdwg.org occurrenceStatus absent present sampling event official","site:docs.gbif.org absence data sampling event dataset guide","Humboldt Core ecological inventories absence data completeness standard official","whitespace visual layout comprehension study primary research","linguistic landscape systematic sampling street segments grid survey signs methodology representativeness","linguistic landscape survey area all signs counted route methodology sampling unit","linguistic landscape database app fields location sign metadata Lingscape methodology","linguistic landscape map no sign absence survey corpus","\"Examining social class and multilingualism through the Linguistic Landscape\" PDF","site:w3.org/WAI/WCAG21/Understanding use of color not sole visual means official","site:w3.org/WAI/WCAG22/Understanding non-text contrast graphical objects official","\"Effect of White-Space Variations on Learning\" 128 university students","site:pmc.ncbi.nlm.nih.gov whitespace visual design comprehension study layout","site:dl.acm.org whitespace readability visual search experiment layout","site:frontiersin.org whitespace visual layout comprehension experiment","negative space visualization comprehension experiment whitespace study"],"sources":[{"source_id":"S1","title":"Examining social class and multilingualism through the Linguistic Landscape: A methodological proposal","publisher":"John Benjamins Publishing Company, Linguistic Landscape","url":"https://benjamins.com/catalog/ll.20032.lu","source_class":"PRIMARY_RESEARCH","publication_date":"2021-11-09","accessed_at":"2026-08-03","claims_supported":["Linguistic-landscape research has recognized representativeness and comparability as methodological concerns.","GIS and census-informed sampling have been proposed to improve the representativeness of sign-based surveys.","Research leads conducting bounded linguistic-landscape studies are a plausible adopter class."]},{"source_id":"S2","title":"Photographic Documentation as a Mode of Data Collection in Qualitative Research: A Case of Pilot Testing in Linguistic Landscape","publisher":"Pakistan Journal of Humanities and Social Sciences","url":"https://journals.internationalrasd.org/index.php/pjhss/article/view/37","source_class":"PRIMARY_RESEARCH","publication_date":"2019-09-30","accessed_at":"2026-08-03","claims_supported":["Photographing signboards and analyzing the language on them is described as the usual linguistic-landscape workflow.","A systematic pilot in two parks treated photographic documentation as the primary collection method.","Interpretation requires the social and physical context surrounding public text, supporting retention of contextual records."]},{"source_id":"S3","title":"Lingscape — Citizen science meets linguistic landscaping","publisher":"Lingscape, University of Luxembourg","url":"https://lingscape.app/","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"Undated current page; includes May 2026 project update","accessed_at":"2026-08-03","claims_supported":["Lingscape collects photographs of signs and lettering and displays uploaded photographs on an interactive map.","The platform offers research project accounts, corpus publication, datasets, and analysis views.","Its stated objectives include analyzing the diversity and dynamics of public writing, showing an identifiable institutional adopter ecosystem.","The public description is positive-record centered and does not describe bounded inspected-no-sign records."]},{"source_id":"S4","title":"DeCiphr — AI-assisted linguistic landscape research and visual data analysis","publisher":"DeCiphr / Salem M. Al-Ajmi","url":"https://deciphr.tech/","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["A current linguistic-landscape product supports GPS field capture, custom coding schemes, structured CSV export, audit trails, inter-rater reliability, and analytic memos.","The product demonstrates that a typed schema, provenance, local storage, and read-only analytical exports are technically feasible in this domain.","It tracks ethics status and offers on-device face blurring, illustrating implementable privacy safeguards.","Its documented unit remains the captured image or sign; it does not advertise exhaustive spatial null slots."]},{"source_id":"S5","title":"Properties of hierarchical events in the Humboldt Extension for Ecological Inventories","publisher":"Biodiversity Information Standards (TDWG)","url":"https://eco.tdwg.org/hierarchy/","source_class":"STANDARD","publication_date":"2024-02-28","accessed_at":"2026-08-03","claims_supported":["A formal inventory standard represents nested observations as bounded events with explicit scope, completeness, effort, and parent-child relationships.","The standard permits zero or non-detection inference only where scope was set in advance and reporting is complete.","It warns against retrospectively inventing scope and against inferring absence from incomplete events.","This is a close cross-domain analogue for typed bounded observation slots and conservative absence inference."]},{"source_id":"S6","title":"Freshwater Data Publishing Guide","publisher":"GBIF Secretariat","url":"https://docs.gbif.org/freshwater-data-publishing-guide/en/","source_class":"OFFICIAL_GUIDANCE","publication_date":"2025-02","accessed_at":"2026-08-03","claims_supported":["GBIF distinguishes opportunistic presence-only observations from planned sampling-event data with documented protocols.","Sampling-event records use event identifiers, locations, dates, effort, and occurrenceStatus to indicate present versus not detected.","The guide identifies missing supporting metadata and ambiguities as obstacles to reuse and meta-analysis.","This provides mature adjacent prior art for the proposal's data model, though not its linguistic-landscape visualization."]},{"source_id":"S7","title":"Web Content Accessibility Guidelines (WCAG) 2.1","publisher":"World Wide Web Consortium","url":"https://www.w3.org/TR/WCAG21/","source_class":"STANDARD","publication_date":"2025-05-06","accessed_at":"2026-08-03","claims_supported":["Non-text content requires an equivalent text alternative unless it is purely decorative.","Color cannot be the only visual means used to convey information or distinguish states.","A null-slot atlas therefore needs programmatic text equivalents, sufficient contrast, and cues beyond color or spatial blankness."]},{"source_id":"S8","title":"Do You See What I See? A Qualitative Study Eliciting High-Level Visualization Comprehension","publisher":"ACM CHI 2024; author manuscript hosted by arXiv","url":"https://arxiv.org/abs/2402.15605","source_class":"PRIMARY_RESEARCH","publication_date":"2024-02-23","accessed_at":"2026-08-03","claims_supported":["Viewer comprehension can diverge from a visualization designer's intended message.","Comprehension varies with visual complexity, composition, data, and viewer background.","The study found stated intentions failed to fully match reported patterns in 59% of tested visualizations.","Direct audience testing is therefore necessary before claiming that protected blank cells improve interpretation or avoid unsupported inference."]}],"problem_evidence":{"support":"MODERATE","rationale":"The sign-centered collection model visibly exists: research and products organize records around photographed signs, while methodological work seeks better representativeness. Mature inventory standards in another field show why absence cannot be inferred without predefined scope, effort, and completeness. However, no direct source documents the prevalence of conflated blank, unvisited, inaccessible, and capture-failure states in linguistic-landscape corpora, or measures resulting analytical errors. The exact problem is plausible and consequential but not yet demonstrated in the target domain.","source_ids":["S1","S2","S3","S4","S5","S6"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Lingscape, its hosted projects, and DeCiphr identify active linguistic-landscape researchers and platform operators who collect, structure, map, audit, and publish sign data. Lu and colleagues express a methodological need for more representative spatial sampling. These are credible adopters or authorizers, but none explicitly requests a bounded-null-slot feature or commits resources to one.","source_ids":["S1","S3","S4"]},"prior_art":{"proximity":"ADJACENT_PRIOR_ART","closest_analogues":[{"name":"TDWG Humboldt Extension hierarchical inventory events","similarity":"Predefines bounded sampling scope, nested observation events, completeness, effort, and rules governing when an unreported target may be treated as undetected.","remaining_difference":"It is a biodiversity data standard, not a linguistic-landscape slot grammar or a figure-ground atlas, and it can reconstitute zeros without rendering them as protected empty cells.","source_ids":["S5"]},{"name":"GBIF sampling-event data with occurrenceStatus","similarity":"Separates planned survey events from opportunistic presence-only observations and associates present/not-detected status with protocol, time, place, and effort.","remaining_difference":"It lacks the proposed six-state linguistic-landscape vocabulary and does not test a quiet null-cell display against a text-filled matrix.","source_ids":["S6"]},{"name":"DeCiphr structured linguistic-landscape workflow","similarity":"Provides GPS capture, custom coding, audit trails, coder agreement, ethics tracking, contextual memos, and structured exports for linguistic-landscape research.","remaining_difference":"Its first-party documentation remains image/sign centered and does not describe mandatory records for inspected locations where no in-scope sign exists.","source_ids":["S4"]},{"name":"Lingscape photo corpus and interactive map","similarity":"Supports institutional projects, geolocated sign photographs, corpus publication, and visual exploration of linguistic landscapes.","remaining_difference":"The public workflow displays contributed positive records rather than a fixed site-slot matrix with typed non-present states.","source_ids":["S3"]}],"distinctive_claim_remaining":"Holding the validated typed slot records, scope, and metadata constant, a repeated atlas that renders inspected-no-sign as a quiet bounded cell will reduce errors or time in identifying the site-level allocation of present signs relative to both a sign-centered baseline and an equally complete text-filled matrix, without increasing slot-state errors, provenance failures, accessibility failures, or unsupported claims of intentional omission. The claim fails if the text-filled matrix performs equivalently without greater crowding or if the quiet cells induce extra inference errors.","confidence":"MODERATE"},"implementation_evidence":{"support":"MODERATE","rationale":"Existing linguistic-landscape software already supplies most needed primitives: custom schemas, GPS records, audit trails, coder-agreement calculations, local storage, ethics fields, and tabular exports. Event-based inventory standards demonstrate a workable relational model for bounded observations and conservative non-detection. A static three-view prototype is technically straightforward. Remaining feasibility gaps concern whether legacy field notes can validate negative observations, whether a common slot grammar fits real sites, and whether spatial/edge cues remain understandable and accessible.","source_ids":["S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":3,"rationale":"Could prevent invalid absence inferences and improve spatial interpretation in bounded surveys, but target-domain prevalence and realized harm are unmeasured and applicability is limited to studies with defensible repeated slots.","source_ids":["S1","S5","S6"]},"stakeholder_pull":{"score":3,"rationale":"Active platforms and research projects visibly seek systematic spatial collection, coding, auditability, and representativeness, but no source expresses demand for this exact feature.","source_ids":["S1","S3","S4"]},"incremental_advantage":{"score":2,"rationale":"Typed scope, sampling events, non-detection, and audit trails are established adjacent practices. Only the protected-null visualization remains incremental, and its advantage over explicit labels has no direct evidence.","source_ids":["S4","S5","S6","S8"]},"distinctiveness_plausibility":{"score":3,"rationale":"No searched linguistic-landscape product or study disclosed the complete combination of predefined site slots, exhaustive typed non-present states, and quiet null-cell rendering. Close cross-domain standards substantially narrow the distinctiveness to the rendering and domain-specific grammar.","source_ids":["S3","S4","S5","S6"]},"technical_implementability":{"score":4,"rationale":"A read-only matrix, status vocabulary, detail view, provenance links, and accessible table export can be built with commonplace data and interface capabilities already demonstrated by domain tools.","source_ids":["S4","S6","S7"]},"adoption_authority_feasibility":{"score":4,"rationale":"A research lead can authorize a retrospective prototype using already permitted records, while site access and new collection remain under existing stewardship and ethics controls. Platform-level integration would require a separate institutional decision.","source_ids":["S3","S4"]},"evidence_readiness":{"score":2,"rationale":"A small comparator test is well bounded, but no validated site dataset, adopter commitment, retrospective-note adequacy assessment, or target-domain error baseline is externally documented.","source_ids":["S2","S8"]},"safety_net_benefit":{"score":1,"rationale":"The proposal primarily improves research validity and interface legibility; no direct benefit to a safety-net population or essential service is demonstrated.","source_ids":[]},"scalability":{"score":4,"rationale":"Fixed slot schemas, typed states, audit records, and tabular exports can be reused across sites, although each study must validate its site grammar rather than impose a universal one.","source_ids":["S4","S5","S6"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"One authorized-site data audit; preregistration; two independent coders; three read-only comparator views; accessibility checks; recruitment and counterbalanced sessions with six analysts; scoring and short report.","confidence":"MODERATE","assumptions":["160-320 loaded labor hours across a methodologist, designer/developer, coders, and analyst","Loaded labor valued at approximately $50-$100 per hour","Participant recruitment, incentives, and administration total $1,000-$3,000","Existing photographs and field notes are available without licensing, travel, or new collection costs"],"source_ids":["S4","S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Production-ready schema and authoring workflow, slot-grammar administration, responsive atlas and detail views, accessible table export, audit/version controls, documentation, security review, and integration with one existing research platform.","confidence":"LOW","assumptions":["Approximately 0.5-1.5 FTE-years distributed among engineering, design, accessibility, research methods, and QA","Existing authentication, storage, image management, and export infrastructure can be reused","No automated computer-vision inference of absence is included"],"source_ids":["S3","S4","S7"]},"operational_launch":{"band_2026_usd":"50K_TO_250K","scope":"Prospective three-to-five-site pilot, protocol development, ethics and site-access administration, fieldworker training, dual coding and adjudication, analyst evaluation, revisions, and deployment support.","confidence":"LOW","assumptions":["600-1,800 staff hours including fieldwork and quality assurance","Sites are local or travel is modest","No restricted-area access, specialized capture hardware, or large public recruitment campaign","Unresolved slots are excluded rather than investigated through costly recollection"],"source_ids":["S2","S5","S6"]},"annual_recurring":{"band_2026_usd":"10K_TO_50K","scope":"Hosting, monitoring, accessibility regression testing, schema/version maintenance, user support, documentation, and minor fixes for one institutional deployment; new field campaigns excluded.","confidence":"LOW","assumptions":["0.1-0.3 FTE ongoing technical and methodological support","Image-storage and traffic volumes remain research-scale","Underlying platform services continue to exist"],"source_ids":["S3","S4"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"UNCERTAIN","reason":"Sources verify sign-centered photographic workflows and analogous absence-inference hazards, but none measures conflated non-present states or their consequences in linguistic-landscape corpora.","source_ids":["S1","S2","S3","S4","S5","S6"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"University-linked Lingscape projects, linguistic-landscape researchers, and DeCiphr's defined research audience constitute identifiable credible adopters; exact feature pull remains unverified.","source_ids":["S1","S3","S4"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim isolates the quiet bounded-null rendering against an equally complete text-filled matrix and a sign-centered baseline, with shared records and observable error, time, provenance, and overinterpretation outcomes.","source_ids":["S5","S6","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"One authorized site, at most six preregistered slots, two coders, three read-only views, and six analysts form a bounded comparator exercise with explicit exclusion and stopping rules.","source_ids":["S4","S7","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The first step can use existing authorized records without new access or capture; provenance is preserved, unverified absences are prohibited, and accessible text equivalents are feasible. Any privacy or access concern triggers rollback to original records.","source_ids":["S4","S7"]},"credible_cost_scope_and_range":{"status":"UNCERTAIN","reason":"Scopes and labor assumptions are bounded, but no direct 2026 wage, integration, recruitment, or institutional overhead evidence was collected; estimates are resource-equivalent planning bands rather than vendor quotes.","source_ids":[]}},"next_evidence_step":"Secure one previously authorized site whose route notes can independently establish no more than six slots. Before viewing sign outcomes, preregister the slot grammar, scope, inspection-time semantics, and scoring rubric. Have two coders independently assign present, inspected-no-sign, inaccessible, not-inspected, out-of-scope, or capture-failure; exclude every unresolved disagreement. Generate from the identical validated table: (A) a sign-centered list/map, (B) a complete text-filled matrix, and (C) the bounded null-slot atlas with keyboard-accessible details and a tabular equivalent. Six eligible analysts see all views in counterbalanced order and answer blinded tasks covering slot-state classification, observable allocation, provenance recovery, uncertainty, and prohibited intent inference; record completion time. Advance only if C has at least a 20-percentage-point lower allocation-error rate than both A and B, adds no more than one total slot-state error relative to B, produces zero excess unsupported-intent claims, and permits provenance recovery on every scored item. Falsify the negative-space advantage if B is within 10 percentage points of C on allocation accuracy with no material time or crowding disadvantage, or if C increases state confusion, provenance failures, accessibility failures, or unsupported intent claims. Separately audit at least three completed target-domain projects for whether their records currently distinguish inspected-no-sign from missing observation; absence of such conflation would falsify the stated problem for the intended adopter class.","blocking_evidence":["No target-domain prevalence estimate shows how often linguistic-landscape corpora conflate observed absence with unvisited, inaccessible, out-of-scope, or failed capture states.","No named adopter has requested, authorized, or funded the null-slot feature.","No live comparator evidence shows that quiet cells outperform an equally complete text-filled matrix.","It is unknown whether retrospective route notes are sufficiently detailed to validate negative observations without reconstruction.","The risk that bounded blanks imply expected signage or institutional intent has not been measured with analysts.","The cross-site validity of a repeated entrance/orientation/transaction/restriction/exit grammar is unknown.","2026 cost ranges lack direct wage, vendor, recruitment, and institutional-overhead evidence."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search found close standards and practices for scoped sampling events, explicit non-detection, completeness metadata, sign-photo corpora, GIS mapping, custom coding, audit trails, and accessible visualization. It did not find the complete linguistic-landscape combination of a predefined site-slot grammar, exhaustive typed non-present states, and protected quiet-cell atlas. This is only a bounded-search contrast: 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":false,"progress_targets":["Obtain a partner-authorized site dataset with prospectively defined or independently verifiable slot inspections; do not reconstruct absences from positive records.","Measure target-domain prevalence by auditing at least three completed linguistic-landscape projects for typed observed-no-sign versus missingness.","Obtain a written adopter or research-lead statement confirming the workflow problem and authority to test the feature.","Preregister the three-view comparison, scoring rubric, allocation-accuracy threshold, non-inferiority limits for state accuracy and provenance, and unsupported-inference falsifier.","Demonstrate accessible keyboard and text/table equivalents that do not rely on blankness, edge patterns, or color alone.","Run the six-analyst counterbalanced test and report all errors, timing, provenance failures, unsupported claims, exclusions, and negative results.","Replace resource-equivalent cost assumptions with locally validated labor, integration, recruitment, overhead, and maintenance estimates."],"reason":"Bounded web research establishes credible adjacent standards, feasible implementation primitives, and an adopter ecosystem, but it cannot determine the target-domain prevalence of the problem or whether protected null cells outperform complete labels without causing overinterpretation or accessibility errors. Those decisive facts require proprietary field records and live analyst testing; under the evaluation rule, this requires an empirical-research stop."},"proposal_index":3}