Standards and precedents
How this archive is built to scholarly practice
The fieldwork layer is an oral-history archive with a dataset attached. This page names the real resources it is modeled on, states what it copies from each and where it still falls short, and audits it against the standards those resources follow. Nothing here is claimed as finished that is not.
The record model
One interview record equals one consented, coded contributor's recorded interview under the project protocol, plus that contributor's completed rating card, described with the metadata below and citable on its own.
Why this archive exists
The gapTwo kinds of resource already exist for color and culture. The World Color Survey and its descendants record how languages name color; the cross-national psychology studies record how people rate color on affective scales. Neither records what a color means to a particular person in their own words, where it is forbidden, what idiom carries it, and how their grandparents read it, tied to that same person's rating on a standard card. That is the unit this archive collects, and it is why the fieldwork layer is structured as an oral-history archive with a dataset attached rather than as a survey alone.
The precedents
Six real resourcesEach row is a resource scholars in the field already cite. The third column is what this archive takes from it; the fourth is the honest difference.
| Resource | What it collects | What this archive copies | The gap |
|---|---|---|---|
| Densho Digital Repository Oral history | Video oral histories on the Japanese American incarceration, with photographs and documents. Every object carries a persistent identifier, interviews are segmented so a single passage can be cited, and each record states its rights tier. (Densho, 2024) | A persistent identifier on every interview record (CTP-CN-01 and so on), a fixed citation form naming the contributor, the interviewer, the date, the identifier, and the collection, and a rights line on every record. | Densho publishes full segmented video; this archive publishes structured excerpts and keeps the recordings in a private tier until written attribution is confirmed. |
| PARADISEC Field recordings of endangered languages | Depositor-driven collections at three levels, collection, item, and file, with ISO 639-3 language codes and technical metadata on every item, and validated spreadsheet self-deposit. (PARADISEC, 2024) | The three-level structure (project, contributor record, recording and card), ISO 639-3 language tags on every interview record, and a deposit path for interviews contributed under the protocol. | PARADISEC stores the primary media files openly where the depositor allows; this archive's media stay private for now. |
| ELAR Sensitive language recordings | Recordings with graded access set per file by the depositor: open, registered users, or restricted. (Endangered Languages Archive, 2024) | A graded model: the public tier is codes only, with quotes and ratings; the private tier holds the name key, the recordings, and the full transcripts until each contributor confirms attribution. | ELAR runs real-name registration for its middle tier; this archive has only an open tier and a private one. |
| The World Color Survey Color naming across languages | Names for 330 Munsell chips from speakers of 110 unwritten languages, plus the best example of each term, with the raw data archived openly at Berkeley. (Kay et al., 2009) | A fixed stimulus set (six named colors, one canonical card) so that answers are comparable across contributors, and open data files published beside the pages. | The WCS records naming, not meaning. It has no testimony, no idioms, no prohibitions, and no generational data. That is the hole this archive collects into. |
| Cross-national semantic-differential studies Cross-cultural psychology | Bipolar ratings of colors in 23 cultures (Adams and Osgood, 1973) and color-emotion associations in 30 nations (Jonauskaite and colleagues, 2020), at national scale. (Adams & Osgood, 1973) | The five-scale rating card is an instrument in this line, so its columns can be read against the published national means. | Those studies have no interview layer: no quotes, no rules, no stories, and no original-language material. This archive pairs every rating with all four. |
| ICPSR and Harvard Dataverse Social-science data archives | A study is its data files plus documentation (codebook, questionnaire, method) plus study-level metadata, with a DOI and an auto-generated citation. | ratings.csv ships with a codebook and the questionnaire on the protocol page, and the dataset has its own citation. | No DOI yet. A Zenodo deposit of the wave 1 dataset is the planned next step, and the DOI will then appear on the how-to-cite page. |
Standards audit
6 in place, 4 partly in place, 1 planned, 1 to confirmThe archive is measured against the standards the precedents follow. A status of Planned or To confirm is a statement of work not yet done, published so a reader can see exactly what the archive is and is not yet.
| Standard | What it requires | Status | Where the archive stands |
|---|---|---|---|
| Oral History Association, Principles and Best Practices (2018) (Oral History Association, 2018) | Informed consent before recording; the narrator's control over attribution; a description of how the interview is preserved and who can access it. | Partly in place | Consent was given on the recording in every wave 1 session (recording, attribution choice, photo use). Written confirmation of each contributor's attribution is being collected; until it lands, every contributor appears by code and no recording is published. |
| Dublin Core (DCMI Metadata Terms) (DCMI Usage Board, 2020) | Each record expressed in a standard element set so a larger archive can harvest it. | In place | Color records, case records, and interview records each carry a Dublin Core block on the page and in the data export. |
| Persistent identifiers | A stable, never-reused identifier on every record. | In place | GCMD-NNN for color records, GCMD-CNN for case records, CTP-<code> for interview records. An identifier is assigned once and never reassigned. |
| Fixed citation form (the Densho model) (Densho, 2024) | A copy-ready citation on every record naming narrator, interviewer, date, identifier, and collection. | In place | Every interview record ends with a Cite this interview block in that form; the how-to-cite page documents it. |
| Language tags (ISO 639-3) (PARADISEC, 2024) | The language of every recording stated in a standard code. | In place | cmn (Mandarin Chinese) and eng (English) on the interview records. |
| Graded access (the ELAR model) (Endangered Languages Archive, 2024) | Sensitive material held at a stricter tier than the public one, with the tier stated. | In place | Public tier: codes, quotes, ratings, metadata. Private tier: name key, recordings, card scans, full transcripts. The tier is stated on every interview record. |
| Technical metadata (PBCore-style) | Format, duration, device, and file details for every recording. | Partly in place | Format and duration are recorded. Device and file specifications are logged privately and not yet listed on the records. |
| Codebook and questionnaire (the DDI and ICPSR model) | Every variable in the dataset defined, with question wording, values, and missing codes; the instrument published. | In place | The protocol page carries the questionnaire and the codebook for ratings.csv. |
| FAIR data principles (Wilkinson et al., 2016) | Findable, accessible, interoperable, reusable. | Partly in place | Findable: stable URLs and identifiers. Accessible: open HTML, CSV, and JSON. Interoperable: Dublin Core and flat CSV with a codebook. Reusable: CC BY 4.0 on record text and data. Missing: a DOI and a machine-readable dataset description. |
| DOI via Zenodo | A persistent identifier for the dataset that resolves independently of this site. | Planned | Deposit the wave 1 dataset (fieldwork.json, ratings.csv, the codebook) on Zenodo once written attributions are in. The DOI then goes on the citation guide. |
| Faculty advisor review | A named scholar who has reviewed the records and the protocol. | To confirm | No advisor is secured. Outreach is in progress. Nothing on this site names an advisor until one has agreed in writing. |
| Contributor credit (CRediT roles) | Every person who did work on the resource named with the role they played. | Partly in place | The curator currently holds every role. Wave 2 interviewers will be credited as Investigation; contributors choose their own attribution. |
Identifier scheme for interview records
Stable IDs| Part | Meaning |
|---|---|
CTP | The Color Testimony Project, the fieldwork collection |
CN, CA, US, ... | Two-letter background code assigned in the private log; never reused, even if an interview falls through |
01, 02, ... | Sequence within that background code |
An identifier is minted once and never reused, even if an interview falls through, which is why sequence numbers can skip. Color and case records keep their own scheme (GCMD-NNN, GCMD-CNN), documented on the how-to-cite page.
The Dublin Core expression of an interview record
Harvestable metadataEvery interview record carries these elements on its page, in this order, so a larger archive or an aggregator can ingest the records without reading the prose.
| Element | Value |
|---|---|
| Identifier | CTP-<code>, permanent |
| Title | Interview with contributor <code> |
| Type | Oral-history interview with rating card (Sound or MovingImage, plus Dataset) |
| Creator | Chloe Chen, interviewer and curator |
| Contributor | The contributor, by code until written attribution is confirmed |
| Date | The wave window; exact dates are held in the private log |
| Language | ISO 639-3 codes of the recording |
| Format | Recorded interview, with duration; rating card as paper or digital form |
| Coverage | The contributor's cultural background and places of upbringing, as stated |
| Subject | Color meaning; the deep-dive colors of the session |
| Relation | The six color records the session feeds; the wave dataset |
| Rights | Record text CC BY 4.0; recording and transcript private tier; consent on the recording |
| Provenance | Collected by the curator under the project protocol; cleaning flags stated on the record |
The citation form for an interview
The Densho modelNarrator, interviewer, date, identifier, collection. The example below is what every interview record's Cite this interview block produces.
Contributor CN-01, interview by Chloe Chen, July to August 2026 (CTP-CN-01). The Color Testimony Project, wave 1, Global Color Meaning Database (Suse), 2026. Accessed [date]. https://globalcolormeaning.org/fieldwork/cn-01.html.
Sources on this page
8 references- Densho (2024). Densho Digital Repository: Using the Repository (identifiers, citation, and rights). Densho. reference
https://ddr.densho.org/using/ - PARADISEC (2024). Pacific and Regional Archive for Digital Sources in Endangered Cultures. PARADISEC. reference
https://www.paradisec.org.au/ - Endangered Languages Archive (2024). Endangered Languages Archive (ELAR). Berlin-Brandenburg Academy of Sciences and Humanities. reference
https://www.elararchive.org/ - Kay, P., Berlin, B., Maffi, L., Merrifield, W. R., & Cook, R. (2009). The World Color Survey. CSLI Publications. ISBN 9781575864150. book
https://linguistics.berkeley.edu/wcs/ - Adams, F. M., & Osgood, C. E. (1973). A Cross-Cultural Study of the Affective Meanings of Color. Journal of Cross-Cultural Psychology 4(2), 135-156. peer reviewed
https://doi.org/10.1177/002202217300400201 - Oral History Association (2018). Principles and Best Practices for Oral History. Oral History Association. standard
https://oralhistory.org/principles-and-best-practices-revised-2018/ - DCMI Usage Board (2020). DCMI Metadata Terms. Dublin Core Metadata Initiative. standard
https://www.dublincore.org/specifications/dublin-core/dcmi-terms/ - Wilkinson, M. D., Dumontier, M., Aalbersberg, I. J., et al. (2016). The FAIR Guiding Principles for Scientific Data Management and Stewardship. Scientific Data 3, 160018. peer reviewed
https://doi.org/10.1038/sdata.2016.18