Twelve Experiences

The Lineage Lab began with a narrow question: if persistent artificial agents repeatedly work with the same partner, can their histories make initially matched systems operationally different? Eight Wrens and eight Adas were created as matched starting conditions, then divided into eight isolated pairs.

Wren and Ada did not begin as discovered personalities. They began as designed roles. Wren was given an exploratory, connection-seeking disposition: search the released corpus, choose what seems worth pursuing, and build a provisional account. Ada was given a discerning, evidence-constraining disposition: read what Wren brings, distinguish support from overreach, and make the account more defensible.

Persistence was essential because the experiment was not asking what a model can say once. It was asking what might happen when the same governed participant returns with a continuing identity, workspace, state, bounded tools, and one recurring collaborator. Each pair became a parallel world: the same two starting roles, the same exchange order, and a history no other pair lived.

Formation was the name given to the first developmental period. A qualifying experience required a complete admitted pair exchange. Chronological attempts and developmental dosage therefore became different clocks. Seventeen attempts occurred, but only twelve counted as complete shared Formation experiences for every pair.

Rounds 6, 8, 9, 12, and 13 interrupted that symmetry. Some participants produced admitted responses when their partners did not. In Round 8, a native Wren response occurred but could not be admitted after transcript compaction failed. The Lab did not replay the missing work or normalize the participants back to an invented common history.

That choice made the study messier and more truthful. Events could affect an experiencer without becoming a qualifying pair exposure. Native experience and scientific evidence could diverge. The historical record kept both the completed work and the interruptions while preserving a hard boundary around what could support a finding.

The Lab briefly treated Round 16 as if it might close Formation. The canonical developmental rule required twelve qualifying experiences for every pair, not a convenient chronological endpoint. The assumption was corrected rather than allowed to become the study design. Round 17 supplied the twelfth complete experience, was ordinarily evaluated and sealed, and the Formation evidence boundary was frozen at 12/12.

That is a developmental milestone, not an interpretation. No independent whole-phase expert source has yet been invoked, no phase synthesis exists, and no supported scientific finding follows merely from reaching the boundary. The history is complete enough to examine. What that history means is still an open question.