The Refaluwasch Journal of Knowledge and Culture (RJKC) is an Indigenous-led scholarly and cultural initiative dedicated to research, documentation, cultural preservation, and the responsible stewardship of Indigenous knowledge. Its formal institutional and legal framework is currently in development.
A responsibly governed digital cultural archive for Refaluwasch and broader Micronesian knowledge — documents, photographs, recordings, oral histories, and the relationships between them.
RJKC's digital archive is under development. No archival collections are currently available through this website. Future records will be subject to documented provenance, appropriate permissions, differentiated access controls, and the cultural authority of relevant knowledge holders.
Not yet available.
Cultural governance
RJKC's model keeps two things firmly apart. The first is the access classification — an enforced authorization attribute that determines who may technically reach a record. The second is the descriptive cultural label — a statement about the cultural terms attached to material. A label describes; a classification enforces. They are recorded separately, and neither is inferred from the other.
A third, separate permission governs whether a record may be retrieved or processed by AI systems. Its default is denied. Publication permission is not AI permission, and an administrator's ability to view a record never implies a right to process it with AI. Access classifications and AI permissions are set and changed by named human administrators only, with a logged approver, and cultural review where the material is culturally governed. No automated system may change them.
The enforced vocabulary
These are the classifications RJKC's architecture defines for archival records. They are described here, not implemented — no record exists, and no classification is enforced anywhere on this site.
Four public safeguards stand behind that vocabulary. Community authority: determinations about the release of culturally governed material rest with people and with the relevant community authority — never with an automated process. Restricted, differentiated access: different audiences may reach different records — some are open; some are for a community, a family, or named holders; some require elder or community authority; some are embargoed until a stated date; some are held for preservation with no access for now; and some must not be digitized at all. Provenance: origins, who supplied the material, and what is known as against what remains uncertain travel with every record. Default-deny AI: every record's AI-retrieval permission defaults to denied and is enforced separately from access, and a record classified do not process with AI is never processed.
| # | Access class | What it means |
|---|---|---|
| 1 | Public | Open to anyone, with no authentication. |
| 2 | Community Access | Open to members of the relevant community — for example the Refaluwasch community, or a specific clan, family, or research group. |
| 3 | Registered Researcher Access | Open to named researchers under a registered-access agreement, with identity verified and purpose and use terms recorded. |
| 4 | Restricted | Access limited to named people or groups, with no public exposure. |
| 5 | Culturally Sensitive | Material whose sharing is constrained by cultural protocol — including sacred knowledge, gendered knowledge, ceremonial content, and knowledge with restricted transmission rules. |
| 6 | Elder or Community Approval Required | Release requires explicit approval from named elders or community authorities. |
| 7 | Embargoed Until Date | Scheduled future release at a specific date — for example after a ceremony, or after a narrator's death. |
| 8 | Private | For the depositor or family only; held in the archive for safekeeping, not for sharing. |
| 9 | Preservation Only | Held for preservation, with no access by anyone for the time being; access decisions are deferred. |
| 10 | Do Not Digitize | The material must not be digitized — physical originals or objects that cultural protocol says should not be copied. The record exists as a description only. |
| 11 | Do Not Process With AI | The record, or its digitized form, must not be processed by any AI system — no transcription, no analysis, no retrieval, no training. |
Every record would also carry an AI-retrieval permission, whose default setting is denied. It is not derived from the access classification and is not visible to AI systems. The two are enforced independently: even where AI retrieval is permitted for an audience, a record classified as Do Not Process With AI is never processed.
Classification vocabulary is held as data rather than hard-coded logic, so that community-defined categories can be added later as controlled vocabulary.
Preservation and provenance
Original and master files are kept as they were received. Derivatives — web previews, compressed audio, viewing copies — are generated from the masters and can be regenerated. A derivative never silently replaces an original, and a reconstruction never stands in for the record it was made from.
Records carry where they came from, who supplied them, what is known, what is remembered, what is interpreted, what is disputed, and what remains unknown — so that uncertainty is never presented as certainty.
Descriptive metadata follows established open standards, preservation events are logged in a preservation-metadata style, and records and their permissions can be exported. The institution's records are designed to outlast any particular platform.
Not yet published. Collection names, scope notes, provenance summaries, and responsible-community attributions will be supplied and approved before publication. No collection is described here.
No holdings, record counts, or growth figures are shown. The pilot archive will begin with a small, carefully governed collection.
Designed, not built
A research system that helps people discover what the archive actually holds is part of RJKC's long-term design. It is not built, and it is not part of the first phase.
The intention is that such a system would help people find and cite material the archive contains — with its sources labelled — rather than manufacture knowledge the archive does not contain. It would be built only after RJKC holds a sufficiently mature, verified, permission-governed collection.
Nothing on this page is an Ask the Archive system, and no query is processed.
Not yet built.
