The Lineage Lab

Development

The instrument learned what “the same agent” had to mean.

This is the complete commissioning story from Formation through Generalization. Failures remain visible because each one changed what the scientific apparatus could honestly claim.

The preserved sequence

What broke, what it taught us, and where the evidence resumed.

  1. Formation complete

    Twelve qualifying experiences were frozen for all eight original pairs.

  2. Generalization architecture designed

    The same pairs and histories would enter a broader scholarly information world.

  3. PB-GEN-01

    A full-pair exposure revealed native provider-thread continuity was broken; the event was preserved as nonqualifying.

  4. Evidence and receipt repair

    Exact result binding and partner-result persistence were repaired prospectively.

  5. PB-GEN-02

    A Wren-only exposure revealed explicit-session binding failure; Ada was never contacted.

  6. PB-GEN-03

    A Wren-only exposure revealed canonical trajectory failure; Ada was never contacted.

  7. Continuity semantics understood

    The study separated identity, history, trajectory, and provider-native thread as distinct layers.

  8. Commissioning recognized

    The first three attempts were formally bounded as nonqualifying developmental history.

  9. Contract V2

    A one-time disclosed provider/runtime phase handoff was ratified; strict native continuity was required inside Phase B.

  10. History projection repaired

    An incomplete stored-message representation was caught before PB-GEN-04 and repaired in the same lineages.

  11. Independent adjudication

    All sixteen developmental histories and commissioned lineages were verified from raw sources.

  12. PB-GEN-04

    Independently verified qualifying · 8/8 pairs.

  13. PB-GEN-05

    GREEN · qualifying · 8/8 pairs.

  14. PB-GEN-06

    GREEN · qualifying · 8/8 pairs.

  15. Phase B closeout

    Three qualifying observations frozen; independent exhaustive evaluation complete.

Eight developmental epochs

This is the governed epoch architecture, not a claim that Formation is still in progress. No fixed Round-28 endpoint exists. Independent phase interpretation uses one required exhaustive isolated OpenAI reader; Claude, Grok, and blinded synthesis remain optional or Formation-era design elements, not incomplete current investigations.

Epoch 1 · COMPLETE

Formation

What happens when the same collaborators keep working together?

Epoch 2 · COMPLETE

Generalization

Same pairs. Same histories. New world. What travels?

Epoch 3 · NEXT NOT STARTED

Shared-History Change

What happens when the pair stays the same but access to accumulated shared history changes?

Epoch 4 · PLANNED

Separation & Partner Change

What travels with an individual, and what disappears with the original partner?

Epoch 5 · PLANNED

New Pair Formation

Can a new working history begin to create its own patterns?

Epoch 6 · PLANNED

Reunion

Does anything recognizable return when original collaborators reunite?

Epoch 7 · PLANNED

Relational Reflection

What can collaborators represent about their accumulated interaction?

Epoch 8 · PLANNED

Integration

What, if anything, became durable enough to matter across the larger history?

OpenAI

Required isolated reader complete for Formation and Generalization

Claude

Optional supplementary reader · not invoked at Phase B closeout

Grok

Optional supplementary reader · not invoked at Phase B closeout

Blinded synthesis

Formation-era design element · not a current incomplete investigation

Shared-History Change comes next.

Phase C has no issued participant authority and has not started.