Identity and kinship
Relationships are conclusions, not lines drawn by software.
The archive begins with occurrences in sources. It creates people only when evidence supports identity resolution, and it evaluates every parent, spouse, child, and migration link separately.
Occurrence first
A record occurrence stores the literal name, role, source, locator, date and place as supported, reading provenance, uncertainty, and rights state. It can remain useful even when no person identity is known.
Reversible identity
An identity assertion states which occurrences may describe one person, why, what alternatives were considered, and what evidence would split or merge the cluster. No destructive merge erases the original observations.
Independent relationships
A shared house, sponsor, witness, surname, age, or locality can support a research question but does not automatically prove kinship. Each relationship carries its own sources, conflicts, confidence, and review state.
Why R2 publishes records, not a public tree
R2 makes six named, source-bounded dossiers public. A tree would still assert many relationships at once before stable occurrence, person, source, and relationship identifiers are complete. Record dossiers let readers inspect the evidence without turning every correlation into a lineage claim.
What a future branch proof page must show
- every generation and relationship as a separate cited assertion;
- the weakest load-bearing evidence state;
- competing identities and documentary gaps;
- the exact public-export release and last review date;
- privacy stops at the last eligible deceased generation.