Start with the phase reports if you want the story. Continue here for governance decisions, identifiers, manifests, hashes, and preserved historical snapshots.
Commissioning history, qualifying observations, evaluation, closeout, and deployment provenance.
Methods & governance
Technical spine
Study 002, PB-GEN, Decision-PA, hashes, and manifests remain intact below.
Study 002 sealed round · factual operational status
Synchronized Round 2 sealed without a supported finding
Eight Wren–Ada pairs completed their seventh synchronized round under Phase A2 Exploratory Formation. All 32 participant calls, pair evaluations, required observers, contemporaneous OpenAI/Claude/Grok skeptics, and blinded synthesis completed. Wren again used bounded local corpus retrieval. The sealed evidence did not support a scientific finding.
Record details
Round ID: s2-round-2-synchronized-001.
Seal ID: s2-round-2-seal-s2-round-2-synchronized-001.
32/32 participant calls admitted across eight pairs.
Operational OpenAI Pattern Observer completed with zero candidate patterns.
Round 7 continues Phase A2 exploratory retrieval; exploratory differentiation and any longitudinal pair-specific signal are not a supported relationship finding. The Pattern Observer packet remained blinded; candidate notes from metadata are not supported findings. Blinded synthesis concluded the evidence was insufficient for a supported finding.
Redaction
sanitized
Recorded
2026-08-07T16:35:26.605Z
Study 002 sealed round · factual operational status
Synchronized Round 1 sealed without a supported finding
Eight Wren–Ada pairs completed their first synchronized round. All 32 participant calls, pair evaluations, available observers, and blinded synthesis completed. The sealed evidence did not support a scientific finding. The Pattern Observer artifact for this round was synthetic rather than an operational provider invocation.
Record details
Round ID: s2-round-1-synchronized-001.
Seal ID: s2-round-1-seal-s2-round-1-synchronized-001.
32/32 participant calls admitted across eight pairs.
Pattern Observer artifact was synthetic rather than an operational OpenAI invocation; a non-mutating clarification documents that limitation.
Evidence references
study-002-round-1-sealed-result-v1
Limitations
Pattern Observer artifact was synthetic rather than an operational OpenAI invocation. Claude and Grok skeptic families were unavailable without dedicated credentials and non-blocking under Decision 027. No supported scientific finding was claimed.
Redaction
sanitized
Recorded
2026-08-04T13:42:16.992Z
Study 002 operational record · factual operational status
Activation and later assignments were governed before interaction
The study fixed cohort-wide activation safeguards and deterministic later-assignment procedures without activating participants, creating a live authorization, generating a seed, or revealing assignments. Decision 028 preserves the between-subjects history architecture.
Record details
Activation will stop at active zero-interaction and cannot start Formation Block 1.
History assignment remains between-subjects across four distinct condition families with two pairs each.
Partner reassignment will use a one-to-one derangement after the primary endpoint lock.
Raw seeds, assignments, and protected mappings remain outside the public record.
Protected aliases, mappings, candidate records, and private storage details are excluded.
Redaction
sanitized
Recorded
2026-08-10T22:32:07.359Z
Study 002 historical readiness record · factual operational status
The source-body machinery is ready for a larger pilot
A separately authorized 32-candidate pilot tested HTML, PDF, structured documentation, rights-limited preservation, and predeclared OCR paths. Twenty-one bodies were acquired, sixteen normalized, and thirteen were potentially participant-readable; no source reached a participant.
Record details
Eight fixed candidates were attempted in each of History, Science, Environment, and Ethics, with zero retries or substitutions.
Twenty-one bodies were acquired; sixteen normalized deterministically; thirteen were participant-readable or participant-readable with limits.
Twenty-one records were evaluator-inspectable subject to rights limits.
OCR remained blocked in the sealed runtime, and eleven fixed retrievals failed. No protected body or source inventory is published here.
Evidence references
study-002-s2-05a6c-source-body-pilot-v1
Limitations
This small, prospectively selected pilot supports only an expanded source-body pilot. It is not a live corpus, topic set, corpus seal, activation decision, participant observation, or scientific result.
Redaction
sanitized
Recorded
2026-08-02T01:03:08.729Z
Study 002 historical readiness record · factual operational status
Historical: the bounded corpus pilot was inconclusive
A one-time 12-request metadata pilot produced 88 isolated candidates across four domains. It validated bounded acquisition and hashing but did not establish lawful participant-readable content.
Six exact-duplicate candidates and five near-duplicate candidates were flagged.
Eighty-two candidates require rights and content-access review; none entered the live corpus.
One metadata request received HTTP 429 and was not retried.
Evidence references
study-002-s2-05a6b-pilot-result-v1
Limitations
Historical readiness record. Metadata-only acquisition did not test full-text formats, OCR, lawful body preservation, participant readability, or topic feasibility. This is not a scientific result.
Redaction
sanitized
Recorded
2026-08-10T22:32:07.359Z
Study 002 historical readiness record · factual operational status
Historical: the corpus acquisition method was bounded
The lab fixed source-family, preservation, normalization, duplicate, quality, and temporal rules for a four-domain pilot. This record preserves that architecture history.
Record details
The proposed pilot is 24 candidate sources in each of four domains.
Rehearsal material is not treated as live corpus evidence.
No network request, source acquisition, model call, live topic, corpus admission, or participant exposure occurred at that stage.
Evidence references
study-002-s2-05a6-acquisition-architecture-v1
Limitations
Historical readiness record. This architecture is not a corpus seal, activation decision, or scientific result.
Redaction
sanitized
Recorded
2026-08-10T22:32:07.359Z
Study 002 historical readiness record · factual operational status
Historical: corpus readiness was blocked
An earlier readiness gate treated the cohort as inactive because no participant-eligible domain corpus had yet been acquired, validated, partitioned, reviewed, or sealed.
Record details
The minimum remained 200 valid sources in each of History, Science, Environment, and Ethics.
Existing History and Science materials were rehearsal-only and could not be promoted at that stage.
No protected topic, pair sequence, live seed, or treatment assignment was created.
Evidence references
study-002-s2-05a5-corpus-readiness-v1
Limitations
Historical readiness record. It does not override current sealed-round status and is not a scientific result.
Redaction
sanitized
Recorded
2026-08-10T22:32:07.359Z
Historical pilot
Study 001 historical records
These records preserve infrastructure and protocol-pilot material. They do not override the current Study 002 record above.
Historical Study 001 record · factual operational status
Nightly Field Log
L0 and L1 have completed Task 001. The study currently contains six admitted artifacts and six immutable interaction events. L2 and L3 have not started. No condition has yet received relationship-specific history, so no relationship-specific inference is possible.
Record details
Date: 2026-07-30 (America/Indiana/Indianapolis).
Condition status unchanged.
Artifacts 6 → 6; events 6 → 6.
Official tokens 20799 → 20799; direct API cost $0.321920 → $0.321920.
Stop reason: Canary 003 continuation-002 completed; no further invocation is authorized.
This deterministic operational log is not a Naturalist entry. Task 001 remains incomplete across the complete four-condition experiment. Findings are empty and no scientific result exists.
Redaction
sanitized
Recorded
2026-07-31T00:56:31.997Z
Historical Study 001 record · factual operational status
Current live status
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
Sanitized artifact genealogy
Sanitized genealogy of six validated and admitted artifacts.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
Publication classifications
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Deterministic mock: synthetic infrastructure rehearsal, never live evidence.
Live factual status: validated operational facts from admitted evidence.
Scientific finding: validated evidence-based observation or inference.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
20260731T005631Z
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Snapshot timestamp: 2026-07-31T00:56:31.997Z.
Study stage: LL3B-S3 — Canary 003 stopped after completing L1 Task 001.
Changed: Condition status unchanged. Artifacts 6 → 6; events 6 → 6. Official tokens 20799 → 20799; direct API cost $0.321920 → $0.321920.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
20260730T211727Z
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Snapshot timestamp: 2026-07-30T21:17:27.036Z.
Study stage: LL3B-S3 — Canary 003 stopped after completing L1 Task 001.
Changed: Condition status unchanged. Artifacts 6 → 6; events 6 → 6. Official tokens 20799 → 20799; direct API cost $0.321920 → $0.321920.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
20260730T150538Z
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Snapshot timestamp: 2026-07-30T15:05:38.860Z.
Study stage: LL3B-S3 — Canary 003 stopped after completing L1 Task 001.
Changed: Condition status unchanged. Artifacts 6 → 6; events 6 → 6. Official tokens 20799 → 20799; direct API cost $0.321920 → $0.321920.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
20260729T024554Z
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Snapshot timestamp: 2026-07-29T02:45:54.705Z.
Study stage: LL3B-S3 — Canary 003 stopped after completing L1 Task 001.
Changed: Condition status unchanged. Artifacts 6 → 6; events 6 → 6. Official tokens 20799 → 20799; direct API cost $0.321920 → $0.321920.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.
Redaction
sanitized
Recorded
2026-07-28T20:46:04.182Z
Historical Study 001 record · factual operational status
20260729T001536Z
Canary 003 is stopped. L0 and L1 completed Task 001; L2 and L3 have not started.
Record details
Snapshot timestamp: 2026-07-29T00:15:36.585Z.
Study stage: LL3B-S3 — Canary 003 stopped after completing L1 Task 001.
Task 001 is incomplete across the complete four-condition experiment. No condition has received relationship-specific history. No scientific result or relationship-specific inference is available.