{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp06_four_proposal_generalization60_20260803","source_assessment_id":"negative_space_design__linguistics_semiotics:P3:v0","cell_id":"negative_space_design__linguistics_semiotics","proposal_index":3,"qualification":"EMPIRICAL_PARTNER_CANDIDATE","criteria":{"specific_differentiated_claim":{"status":"YES","reason":"The remaining claim specifically predicts that protected quiet null cells outperform an equally complete text-filled matrix on allocation accuracy or time without worsening state accuracy, provenance recovery, accessibility, or unsupported inference."},"credible_problem_signal":{"status":"YES","reason":"Target-domain sources establish sign-centered photographic workflows and concern about representative spatial sampling, while mature adjacent inventory standards substantiate the hazard of inferring absence without predefined scope and complete observation records. Prevalence remains uncertain but the external signal is credible."},"identifiable_partner_or_adopter":{"status":"YES","reason":"Linguistic-landscape research leads, university-linked projects, and platform operators such as those maintaining structured sign corpora form a concrete adopter class with authority over protocols, records, and analytical interfaces."},"partner_access_is_necessary":{"status":"YES","reason":"Resolving target-domain prevalence, validating negative observations, and measuring analyst interpretation requires authorized project records, route notes, workflow knowledge, and eligible analysts that public research cannot supply."},"safe_authorized_first_step":{"status":"YES","reason":"The first step uses an already authorized site and read-only derived views, prohibits reconstructed absences, excludes unresolved slots, preserves provenance and accessible equivalents, and provides explicit privacy, authority, halt, and rollback rules."},"bounded_decisive_empirical_design":{"status":"YES","reason":"The design bounds the grammar to six slots, uses two independent coders, identical validated records, three comparator views, six counterbalanced analysts, explicit metrics, advancement thresholds, and falsification conditions. A three-project audit separately tests whether the target problem exists in the intended workflow."},"no_material_negative_gate":{"status":"YES","reason":"No verified pipeline gate is NO: adopter credibility, the incremental claim, bounded evidence step, and safety/authority all pass. Problem prevalence and cost are uncertain, but neither creates a safety stop or defeats the proposed empirical test."},"not_merely_more_research":{"status":"YES","reason":"The next action is an operational partner exercise specifying datasets, authorization, slot limits, coding procedure, comparators, participants, metrics, thresholds, and falsifiers rather than an open-ended request for further research."}},"uncertainty_types":["PROBLEM_PREVALENCE","ADOPTER_PULL","INCREMENTAL_EFFECT","WORKFLOW_FIT","DATA_ACCESS","COST_SCOPE"],"partner_profile":"A linguistic-landscape research lead or institutional platform team controlling a completed, authorized bounded-site survey with route notes detailed enough to verify inspected-no-sign and other non-present states, plus access to eligible analysts and authority to create read-only research views.","required_access":"Authorized photographs, route and field notes, scope definitions, inspection times, access and ethics metadata, coder participation, and six eligible analysts; additionally, auditable records from at least three completed target-domain projects for the prevalence check.","bounded_empirical_test":"Preregister at most six slots for one independently verifiable authorized site; dual-code typed states and exclude unresolved slots; generate a sign-centered baseline, complete text-filled matrix, and accessible null-slot atlas from the identical validated table; run six analysts in counterbalanced order on state, allocation, provenance, uncertainty, inference, and timing tasks; separately audit three completed projects for absence-versus-missingness distinctions.","success_condition":"The atlas achieves at least a 20-percentage-point lower allocation-error rate than both comparators, adds no more than one total state-classification error versus the text matrix, causes zero excess unsupported-intent claims, and permits provenance recovery on every scored item; the project audit also finds material target-domain conflation of observed absence and missingness.","falsification_condition":"Falsify the problem for the intended class if audited projects already distinguish observed-no-sign from missing observation reliably. Falsify the visualization advantage if the text matrix is within 10 percentage points of atlas allocation accuracy without material time or crowding disadvantage, or if the atlas increases state confusion, provenance or accessibility failures, or unsupported intent claims.","rationale":"This is a narrow empirical-partner case: adjacent standards substantially collide with the data-model idea, but a differentiated domain-specific visualization claim remains. External evidence makes the underlying problem credible without establishing its prevalence, and the decisive residual questions require authorized records and live analyst behavior. The proposed partner test is safe, comparator-based, bounded, and explicitly capable of rejecting either the problem premise or the negative-space advantage."}