*SET-U open data & archive

Open data

All 27 *SET-U datasets in full, unpaginated, as published by the live atlas's export contract (schema 2026-08-28.107). Rebuilt from the reviewed migrations whenever they change.

Download the whole atlas as SQLite 5.0 MB · manifest.json

DatasetRowsFiles
entries
Reviewed word dossiers and their editorial claims.
214 JSON · CSV
languages
Language and reconstructed-language identities used by canonical nodes.
92 JSON · CSV
nodes
Canonical attested, reconstructed, and knowledge-boundary lexeme nodes.
1,061 JSON · CSV
edges
Canonical directed historical relationships, with sound, meaning, scholar, and confidence claims.
1,025 JSON · CSV
changes
A discovery-ready change ledger of every reviewed canonical transformation, preserving endpoint forms, languages, relation type, sound history, semantic history, scholar, confidence, chronology, and citation totals.
1,025 JSON · CSV
edge-citations
Source-level support, challenge, or context attached to canonical graph edges.
1,669 JSON · CSV
descendant-closure
Every reviewed root-to-downstream-edge reachability record, with shortest historical depth; join edgeId to edges and edge-citations for the full evidence chain.
9,038 JSON · CSV
families
Every connected component in the reviewed lexical graph, with dossier, node, edge, citation, language, dispute, witness, and truth totals. Historical-context links are excluded.
54 JSON · CSV
family-memberships
The exact assignment of every reviewed modern-word dossier to its graph-derived evidence family.
214 JSON · CSV
dossiers
Reviewed entry-scoped paths that map editorial dossier forms onto canonical nodes.
1,366 JSON · CSV
journeys
Ordered geographic and temporal stages behind dossier journey views, including honest null coordinates and placement confidence.
959 JSON · CSV
timeline-events
Cross-word chronology of reviewed canonical forms and directly cited journey stages with explicit numeric bounds; the two evidence classes remain distinguishable.
1,398 JSON · CSV
journey-citations
Direct source stances, locators, and evidence notes attached to geographic and temporal journey-stage claims.
1,324 JSON · CSV
claims
Entry-scoped reviewed ancestry claims and their canonical-edge mappings.
1,152 JSON · CSV
truths
Ranked word revelations with an explicit root, root gloss, transformation path, unveiling rationale, reveal type, evidence grade, and permanent record URLs.
214 JSON · CSV
truth-citations
Direct source-level support or context attached to each ranked contrarian reveal.
569 JSON · CSV
sources
Bibliographic source register with live usage counts. Linked works retain their own rights and terms.
624 JSON · CSV
attestations
Source-reported and directly inspected lexical witnesses with explicit provenance status, named sources, and permanent audit URLs.
538 JSON · CSV
witness-inspections
Repeatable visual-inspection records for directly inspected primary witnesses, preserving the observed token, context, exact locator, artifact, method, date, role, and claim limit.
46 JSON · CSV
disputes
Competing or unresolved scholarly analyses preserved without false convergence, with permanent audit URLs.
865 JSON · CSV
dispute-citations
Direct source stances, locators, and evidence notes attached to scholarly-dispute positions.
1,602 JSON · CSV
context-links
Sourced historical connections that are explicitly not lexical descent or cognacy.
20 JSON · CSV
context-link-sources
Source-level evidence attached to non-genealogical historical context links.
54 JSON · CSV
dossier-audit
A non-composite evidence audit of every reviewed word across lineage, change claims, witnesses, journeys, disputes, network depth, linked sources, and open research tasks.
214 JSON · CSV
lexical-candidates
Evidence-seeded attested or bound forms that appear as cited relatives in the reviewed graph but do not yet have complete dossiers.
505 JSON · CSV
candidate-resolutions
Source-cited editorial decisions that a graph-visible candidate is a represented variant, merged identity, non-lexeme, or duplicate rather than a separate dossier.
3 JSON · CSV
research-frontier
A live work queue derived from open disputes, uninspected witnesses, knowledge boundaries, low-confidence edges, thin language representation, and evidence-seeded words awaiting complete dossiers.
1,632 JSON · CSV

Reuse

SET-U has not yet declared reuse terms for its original editorial data, so ask before republishing it wholesale. Every cited source work keeps its own rights; quotations are short and attributed. Code that builds this edition is MIT-licensed.

How the files relate

Nodes are historical forms; edges connect a child form to its parent with the sound and meaning change and a confidence grade; edge-citations link each edge to its sources with a locator and stance. Dossiers (entries) follow an editorial path through the graph (dossiers). Full field notes: docs/public-data.md.