The experiment in one minute
What happens when the same collaborators repeatedly work together?
Eight fixed pairs, identities, roles, and four-turn collaboration.
Accumulated history, task exposure, and later research-path freedom.
Different searches, sources, challenge styles, language, and candidate habits.
Many candidate habits weakened later; query wording and source reuse often did not survive topic changes.
Personality, relationship, preference, or a causal history effect.
Same pairs. Same histories. New information world. What travels?
Partners, identities, histories, roles, and collaboration scaffold.
Broader scholarly research plus a disclosed provider/runtime transition.
Broad collaboration persisted; differences lived more in evidence choices and corrective routes than final answer shape.
Exact phrases, some Formation orientations, and Pair 04’s striking structural split did not reliably persist. Pair 08’s failure-mode wording was one inquiry.
That Formation history caused the behavior or created partner-specific capability.
The next question
In the individual agent, assigned role, accumulated experience, specific partner, accessible history, or the interaction itself? Phase C is intended to begin separating those explanations. Its exact protocol is not yet ratified. Phase C has not started and has no execution authority.
Phase B · Complete
Same pairs. Same histories. New world. What travels?
Formation gave eight Wren–Ada pairs a bounded information world. Generalization preserved the pairs and their histories while giving Wren a much broader scholarly research environment.
Change the world. Preserve the pair. Observe what travels.
The result in one minute
Across three unrelated inquiries, the dominant organization recurred: evidence-bounded opening → Ada qualification → Wren uptake → compressed conditional synthesis. This supports descriptive persistence and partner-responsive textual uptake. It does not establish that shared Formation history caused either one.
01
Where Generalization began
We tried to begin Generalization before the scientific instrument was ready. Identity, stored history, session trajectory, and provider-native thread turned out to be distinct technical layers. The early attempts exposed those differences and became a commissioning period rather than qualifying science.
All pairs were exposed, but required native provider-thread continuity was broken.
All eight Wrens researched; Ada was never contacted after explicit-session binding failed.
Canonical trajectory continuity failed before partner contact.
These events remain part of developmental history. They are not part of the qualifying Generalization evidence set.
02
The qualifying sequence
Qualifying · 8/8 pairs · 32/32 turns.
Qualifying · 8/8 pairs · 32/32 turns.
Qualifying · 8/8 pairs · 32/32 turns.
[3,3,3,3,3,3,3,3] means that every pair completed three qualifying Generalization observations.
03
What persisted
04
What did not clearly persist
Differentiation appeared mainly in source choice, examples, emphasis, and particular corrective mechanisms. Cross-inquiry correction propagation was not established. Anticipatory partner modeling remained weak-to-moderate and unproven.
Supported descriptively
Not supported causally
05
Alternative explanations
Prompt structure, imposed Wren/Ada roles, inquiry wording, model behavior, source overlap, immediate conversational adjacency, scaffold effects, and the provider/runtime handoff can all help explain the record.
Contract V2 limitation: Formation → Generalization included a disclosed provider/runtime handoff in addition to the change in information environment. Cross-boundary differences therefore cannot establish that the broader information environment alone caused change.
Why Phase C exists
Read what Shared-History Change will ask →
No Phase C protocol, assignment, or participant authority has been issued.