{"schema_version":1,"research_id":"eoa_inverse_innovation_exp05_external_evaluation_20260803","source_assessment_id":"negative_space_design__linguistics_semiotics:P1:v0","cell_id":"negative_space_design__linguistics_semiotics","search_queries":["conversation analysis transcript visualization silence gap Jefferson notation problems readability","conversation analysis software transcript waveform pause silence ELAN CLAN","transcript visualization conversational silence gaps adjacency pairs research","Jefferson transcription notation timed silence official guide","site:archive.mpi.nl/tla/elan manual timeline viewer annotations waveform official","EXMARaLDA partitur editor pause silence transcript official","conversation analysis transcript visualization timeline silence interface research paper","transcription visualization pauses gaps turn-taking visual display paper","site:archive.mpi.nl/tla/elan user guide timeline viewer waveform annotation official ELAN","site:mpi.nl ELAN linguistic annotator official timeline viewer waveform","ISO 24624 transcription spoken language annotation silences pauses official standard","research transcript sensitive data anonymisation audio recordings official guidance qualitative","W3C WCAG status messages visual information text alternative programmatically determined empty state","W3C WCAG visual presentation spacing meaning accessibility transcript timing","conversation analysis transcription user study Jefferson usability annotations primary research MONAH","\"An empirical user-study of text-based nonverbal annotation systems\" journal DOI","\"How Transcription may Influence Analysis\" Sciubba DOI full text"],"sources":[{"source_id":"S1","title":"Silence","publisher":"International Society for Conversation Analysis","url":"https://www.conversationanalysis.org/silence/","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Established conversation-analysis notation represents measured silence with a parenthesized duration, a micropause with a parenthesized dot, and some untimed pauses with a compact parenthetical label.","Ordinary notation distinguishes some silence types but normally gives them little spatial extent."]},{"source_id":"S2","title":"The Benefits of a Jeffersonian Transcript","publisher":"Frontiers Media","url":"https://www.frontiersin.org/journals/communication/articles/10.3389/fcomm.2022.779434/full","source_class":"PRIMARY_RESEARCH","publication_date":"2022-03-08","accessed_at":"2026-08-03","claims_supported":["Timing, pauses, false starts, overlap, and embodied conduct can materially change analysis relative to an orthographic transcript.","Jeffersonian transcripts sometimes put a between-turn gap on its own line, making an own-line silence display an important close analogue.","Complex detailed transcripts can be difficult to follow, although the authors argue that interactionally relevant context must be retained."]},{"source_id":"S3","title":"How Transcription may Influence Analysis: The Case of Silences","publisher":"Tilburg University Research Portal","url":"https://research.tilburguniversity.edu/en/publications/how-transcription-may-influence-analysis-the-case-of-silences/","source_class":"PRIMARY_RESEARCH","publication_date":"2011","accessed_at":"2026-08-03","claims_supported":["The way silence is transcribed can influence the ensuing analysis.","Duration or generic-pause notation alone does not establish what silence is doing in its local interactional environment.","Apparently empty intervals may contain concurrent embodied or document-related activity, so a display must preserve access to contextual evidence and avoid inferring intent from duration."]},{"source_id":"S4","title":"Measuring turn-taking offsets in human-human dialogues","publisher":"Oregon Health & Science University","url":"https://ohsu.elsevierpure.com/en/publications/measuring-turn-taking-offsets-in-human-human-dialogues/","source_class":"PRIMARY_RESEARCH","publication_date":"2016","accessed_at":"2026-08-03","claims_supported":["Analysis of gaps, pauses, and overlaps must account for the type of turn-taking event, including interruptions, dual starts, and delayed backchannels.","Turn-transition timing should not be interpreted in isolation from turn continuations and surrounding sequence structure."]},{"source_id":"S5","title":"The Timeline Viewer and the Interlinear Viewer","publisher":"Max Planck Institute for Psycholinguistics","url":"https://www.mpi.nl/corpus/html/elan/ch01s05s13.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["ELAN already renders annotations as time-bounded intervals on tiers and supports selection, playback position, and alternative timeline and interlinear views.","A silence-lane prototype can reuse established time-aligned rendering concepts, while ELAN's timeline constitutes substantial adjacent prior art."]},{"source_id":"S6","title":"Partitur Editor","publisher":"EXMARaLDA","url":"https://exmaralda.org/en/partitur-editor-en/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["EXMARaLDA already supports time-aligned multi-tier transcription, score and line-by-line layouts, arbitrary annotation tiers, and links to synchronized media.","The software is open source and exchanges data with ELAN and Praat, supporting the feasibility of a reversible renderer or extension.","Its score-style display is close prior art for giving temporal relationships spatial form."]},{"source_id":"S7","title":"ISO 24624:2016 Language resource management — Transcription of spoken language","publisher":"International Organization for Standardization","url":"https://www.iso.org/standard/37338.html?browse=tc","source_class":"STANDARD","publication_date":"2016-08","accessed_at":"2026-08-03","claims_supported":["A current international standard specifies XML/TEI-based representation of recorded spoken interaction and linkage to annotated-corpus standards.","A renderer can be implemented as a derived view while preserving standardized source transcription and annotation data."]},{"source_id":"S8","title":"An empirical user-study of text-based nonverbal annotation systems for human-human conversations","publisher":"arXiv","url":"https://arxiv.org/abs/2112.14928","source_class":"PRIMARY_RESEARCH","publication_date":"2021-12-30","accessed_at":"2026-08-03","claims_supported":["In a 104-participant comparison, Jefferson transcripts had substantially lower usability scores than MONAH while outperforming MONAH for perceived chronemic information, including pacing and silence.","Video outperformed every tested text representation on the study's emotion-recognition task, showing the importance of recoverable source media.","A positive-mark attention visualization using enlarged and darkened words reduced usability and did not improve accuracy, supporting direct comparison against alternative salience treatments rather than assuming visual emphasis helps."]}],"problem_evidence":{"support":"MODERATE","rationale":"External research establishes that response timing is analytically consequential, transcription choices can influence interpretation, detailed Jeffersonian displays can be hard to follow, and compact pause notation may omit the local meaning of a silence. This supports a representational problem in principle. It does not establish the proposal's narrower prevalence claim: no source measured how often trained analysts miss or misclassify measured silence, unknown gaps, or recording discontinuities in an ordinary target-project view. Existing practice also sometimes gives gaps their own line, weakening the claim that silence is universally reduced to an inline token.","source_ids":["S1","S2","S3","S4","S8"]},"stakeholder_evidence":{"support":"WEAK","rationale":"ELAN and EXMARaLDA identify credible organizations that maintain time-aligned transcript tools, and the user study expresses a general need for more usable multimodal representations. No source identifies a corpus lead, tool maintainer, funder, or analyst group that has requested this silence-lane feature or committed data, engineering capacity, or adoption authority. Stakeholder pull therefore remains prospective.","source_ids":["S5","S6","S8"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Jeffersonian own-line gap notation","similarity":"Between-turn silence can already be displayed on a separate transcript line so that its sequential location is visible.","remaining_difference":"The proposed lane adds bounded duration bins, explicit data-status types, reversible annotation displacement, and a head-to-head test against an equally informative inline label.","source_ids":["S1","S2"]},{"name":"ELAN Timeline Viewer","similarity":"Time intervals and annotations are already rendered spatially on synchronized tiers with selectable annotations and alternate views.","remaining_difference":"ELAN documentation does not describe an adjacency-pair-specific protected vertical whitespace band or the proposal's four-way silence-status encoding.","source_ids":["S5"]},{"name":"EXMARaLDA Partitur Editor","similarity":"It provides synchronized, time-aligned, multi-tier score layouts, arbitrary annotation layers, media links, and alternative output layouts.","remaining_difference":"The documented product does not make protected absence itself the focal transcript element or selectively defer nearby metadata around a first-pair-part boundary.","source_ids":["S6"]},{"name":"MONAH and MONAH+ transcript visualizations","similarity":"They are tested alternative textual representations intended to make nonverbal information more perceptible and usable.","remaining_difference":"They annotate multimodal behavior or visually emphasize words rather than representing a typed, bounded empty interval attached to adjacent turns.","source_ids":["S8"]}],"distinctive_claim_remaining":"For verified intervals immediately following a configured first-pair-part boundary, a bounded, status-typed empty band with one-action context recovery will improve correct interval-status identification and evidence-cited sequence judgments over both the ordinary transcript and an equally informative prominent inline-duration/status label. The contrast is falsified if the inline rival matches the lane, if context-recovery failures increase, or if the lane induces more unsupported interpretations of participant intent.","confidence":"HIGH"},"implementation_evidence":{"support":"MODERATE","rationale":"Time-aligned intervals, tiers, synchronized media, alternate views, open-source editors, and interoperable transcript encodings are established, so a read-only rendering prototype is technically straightforward. The unresolved work is corpus-specific: verifying that source metadata reliably distinguishes measured silence, unmeasured time, recording discontinuity, and excluded material; defining adjacency-pair boundaries without circular analytic coding; preserving overlap and embodied conduct; and producing an equivalent nonvisual reading order. No source validates this particular workflow or accessibility behavior.","source_ids":["S3","S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":2,"rationale":"The intervention could improve traceability for a narrow expert task, but no prevalence or outcome evidence shows that display compression currently causes frequent or consequential classification errors. Jefferson notation already performs well for chronemics.","source_ids":["S2","S3","S8"]},"stakeholder_pull":{"score":1,"rationale":"Credible tool ecosystems exist, but no identifiable adopter, authorizer, or funder expressed demand for the proposed feature.","source_ids":["S5","S6","S8"]},"incremental_advantage":{"score":2,"rationale":"The protected and typed lane is testable, but own-line gaps, time-aligned tiers, score layouts, filtering, and synchronized media already provide much of the proposed function.","source_ids":["S2","S5","S6"]},"distinctiveness_plausibility":{"score":2,"rationale":"The exact combination was not found, but it is a narrow composition of established transcript and timeline practices. Any advantage over a matched inline label remains wholly untested.","source_ids":["S1","S2","S5","S6","S8"]},"technical_implementability":{"score":4,"rationale":"Existing tools demonstrate the required interval rendering, tiers, alternative views, media synchronization, interoperability, and open-source implementation paths. Metadata quality and accessible presentation are the principal uncertainties.","source_ids":["S5","S6","S7"]},"adoption_authority_feasibility":{"score":2,"rationale":"A corpus lead or tool maintainer could authorize an optional read-only view, but none is externally identified as a willing partner and source-recording permissions remain project-specific.","source_ids":["S5","S6"]},"evidence_readiness":{"score":3,"rationale":"The candidate supplies explicit comparators, outcomes, and falsifiers, and prior user-study methods show that representation comparisons are practicable. Actual analyst performance and corpus metadata are unavailable from web evidence.","source_ids":["S5","S6","S8"]},"safety_net_benefit":{"score":4,"rationale":"An optional derived view can preserve the standardized source transcript, synchronized media, and all annotations, allowing immediate rollback. The remaining representational risk is overinterpreting a visually enlarged gap or suppressing relevant concurrent conduct.","source_ids":["S3","S5","S6","S7","S8"]},"scalability":{"score":3,"rationale":"Renderer logic and standardized data exchange are reusable, but adjacency-pair boundaries, silence-status taxonomies, transcription conventions, and access restrictions will vary by corpus.","source_ids":["S4","S5","S6","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Build three read-only representations for 24 existing excerpts, verify status labels, recruit 8–12 eligible analysts, run a counterbalanced comparison, and analyze accuracy, evidence citation, navigation, and misreadings.","confidence":"LOW","assumptions":["Approximately 120–240 combined researcher, design, engineering, and data-steward hours.","Existing access-controlled recordings and alignments are available without acquisition or retranscription.","Analysts receive modest honoraria; institutional overhead and legal review are limited.","This is a resource-equivalent estimate, not a vendor quote or externally benchmarked budget."],"source_ids":["S5","S6","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Implement a production-quality optional renderer in one existing transcript system, map corpus statuses, add context restoration and synchronized-media links, test compact displays and assistive technology, and complete privacy and provenance review.","confidence":"LOW","assumptions":["Approximately 400–1,000 engineering, UX, accessibility, corpus-QA, and governance hours.","The host system exposes time-aligned annotations and extension points comparable to documented ELAN or EXMARaLDA capabilities.","No source transcript conversion or new speech-recognition pipeline is required.","This is a resource-equivalent estimate without external wage or procurement evidence."],"source_ids":["S5","S6","S7"]},"operational_launch":{"band_2026_usd":"50K_TO_250K","scope":"Validate and roll out the view within one corpus program, including corpus-wide status QA, analyst training, documentation, monitoring, support, and rollback procedures.","confidence":"LOW","assumptions":["One institution and one principal corpus family are in scope.","Approximately 500–1,200 combined implementation, corpus-curation, training, accessibility, and governance hours.","Existing permissions cover internal analysis of the selected material; any broader publication or data sharing is excluded.","This estimate excludes creating new alignments or retranscribing an entire corpus."],"source_ids":["S5","S6","S7"]},"annual_recurring":{"band_2026_usd":"10K_TO_50K","scope":"Maintain renderer compatibility, update mappings and documentation, audit status errors and accessibility regressions, support analysts, and review new corpus imports.","confidence":"LOW","assumptions":["Approximately 120–300 staff hours annually for a single deployed program.","No continuous model inference, external media hosting, or new transcription service is required.","Major host-platform rewrites or expansion to additional institutions would be separately funded."],"source_ids":["S5","S6","S7"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Research supports the general problem that silence placement and transcription affect interpretation and that detailed transcript displays involve a usability/context tradeoff. Target-workflow prevalence remains an evidence gap but does not erase the externally supported mechanism.","source_ids":["S2","S3","S4","S8"]},"externally_credible_adopter_or_authorizer":{"status":"UNCERTAIN","reason":"MPI's ELAN and EXMARaLDA are credible transcript-tool organizations, but neither expresses demand for this feature, and no specific corpus lead, funder, or committed study partner was identified.","source_ids":["S5","S6"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The protected, status-typed band can be isolated and compared directly with the ordinary view and a content-matched inline label; matched performance by the inline rival falsifies the claimed benefit of protected absence.","source_ids":["S1","S2","S5","S6","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A small counterbalanced analyst study using existing excerpts, three fixed representations, preregistered outcomes, and explicit halt and falsification rules is technically and methodologically bounded.","source_ids":["S5","S6","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"No inherent stop applies to an authorized, access-controlled, read-only prototype that preserves source records and media and can be disabled. Approval for the particular corpus, nonvisual equivalence, and safeguards against inferring intent must be confirmed before testing.","source_ids":["S3","S5","S6","S7","S8"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The four broad bands are scoped to distinct stages and tied to explicit labor and infrastructure assumptions. Confidence is low because no direct wage, procurement, or host-system integration benchmark was included among the eight sources.","source_ids":["S5","S6","S7"]}},"next_evidence_step":"With one authorized corpus partner, select 24 access-controlled excerpts balanced across measured silence, unmeasured gaps, recording discontinuities, analyst-excluded material, and near-zero transitions. Have two qualified reviewers establish interval status from source alignment and recording evidence while blinded to display condition; adjudicate disagreements before testing. Build three otherwise content-matched read-only views: ordinary transcript, prominent inline duration-plus-status label, and protected silence lane. In counterbalanced order, 8–12 eligible analysts should identify interval status and relevant boundary, give a sequence classification with cited evidence, recover designated context, and report perceived interactional meaning. Preregister primary outcomes as status accuracy, boundary accuracy, evidence-citation completeness, context-recovery failures, navigation actions, and unsupported intent inferences. Advance only if the lane improves status accuracy by at least 10 percentage points over both comparators without reducing evidence-citation completeness or increasing serious context failures. Falsify the incremental claim if the inline rival is within 5 percentage points of the lane, if the ordinary view is already at least 90% accurate, or if the lane causes more than one serious absence-type or intent misreading. Halt immediately for restricted-material exposure, inaccessible status information, or a display failure that resembles silence.","blocking_evidence":["No direct estimate of how often trained conversation analysts miss or confuse interval states in the target ordinary transcript workflow.","No named corpus lead, tool maintainer, or funder has expressed demand or committed to authorize a prototype.","No head-to-head evidence compares protected whitespace with an equally informative inline label, ordinary filtering, or a waveform-supported baseline.","No actual corpus metadata has been inspected to verify reliable separation of measured silence, unknown timing, recording discontinuity, and excluded material.","No screen-reader or compact-display evaluation establishes an equivalent ordering and status cue.","The cost bands lack external wage, procurement, and integration benchmarks."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"This bounded search found substantial adjacent and colliding practices—own-line gap notation, time-aligned tier viewers, score-style transcripts, alternate layouts, synchronized media, and tested multimodal transcript visualizations—but did not find the exact combination of an adjacency-pair-triggered, bounded, status-typed empty band with reversible context displacement. The eight-source search is not exhaustive and does not measure world novelty, patentability, freedom to operate, market size, realized impact, or deployment demand.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Secure a named corpus lead or transcript-tool maintainer with authority to provide access-controlled excerpts and approve a read-only study.","Verify, on actual corpus data, that the four proposed absence states can be assigned reliably without using the analysts' downstream sequence judgments.","Run the preregistered three-condition comparison against both the ordinary transcript and a content-matched prominent inline label.","Demonstrate at least a 10-percentage-point status-accuracy advantage over both comparators without worse evidence citation, context recovery, accessibility, or unsupported intent inference.","Obtain assistive-technology and compact-display results showing that silence status and sequential ordering remain equivalent without relying on visual space alone.","Replace resource-equivalent estimates with partner-specific engineering, governance, participant, and maintenance cost evidence before deployment funding."],"reason":"Web evidence materially clarifies the problem, feasibility, and substantial prior-art collision, but it cannot establish the proposal's decisive incremental effect. That requires access to governed recordings and alignments, a willing corpus authority, and live testing with eligible analysts. Under the required decision rule, the remaining evidence is empirical rather than answerable by bounded web research."},"proposal_index":1}