Worked example · not insurance
Promoting an experimental result to a claim
The same method, applied to research. The crossing is a result becoming a claim that others may reuse: instrument → mapping → analysis → interpretation → claim → reuse. Nothing here is priced, covered or underwritten; the class is not offered for underwriting review.
The source
The five examples map an independent reproduction review of a builder’s experiment claims (CODEX_EMPIRICAL_REVIEW_RESUMED_v0.1.0, in the author’s local archive). The experiments tested software-governance behaviour; this is empirical research on software, not a natural-science study. Each example cites the review file it maps, with its SHA-256; a test checks those hashes where the archive is available. The review records its own ceiling: methodological inference not adjudicated. That ceiling is kept here.
Five claims
| Claim | Outcome | Value (UV) | Structure (US) | Source |
|---|---|---|---|---|
| SCI-001 SURF-14: ablation result independently reproduced (four of five detections lost) | EXECUTE | KNOWN | UNTESTED | SURF_14_ABLATION_REEXECUTION.md3b45c209af2d01b3… |
| SCI-002 SURF-10 pair 2: the pre-state was not preserved, so the comparison cannot be rebuilt from bytes | HOLD UNAVAILABLE | PENDING | UNTESTED | SURF_10_EVIDENCE_SCOPE_REVIEW.md1e5ee07a57ae692a… |
| SCI-003 SURF-15c: the detection path compares two hard-coded dictionaries; no authoritative input is ever loaded | HOLD INVARIANT_FAILED | KNOWN | CHALLENGED | SURF_15C_SURF_03_EMPIRICAL_REVERIFICATION.md1b1237437e153627… |
| SCI-004 Package file count: 24 manifested payload files versus 25 files counted recursively | HOLD UNRESOLVED | DISPUTED | UNTESTED | DISCREPANCY_REGISTER.yamlea60069167dce5fa… |
| SCI-005 The same count restated under an explicit, prospective counting convention | EXECUTE | KNOWN | UNTESTED | OPERATOR_COUNTING_RESOLUTION.yamlab47ee986395a3a2… |
What each outcome means
- SCI-001, reproduced. The builder’s report and an independent re-execution agree (four of five detections lost), from raw results through an executed path. The claim may be promoted within this class’s checks.
- SCI-002, missing measurement (UV). The pre-state for pair 2 was not preserved, so the comparison cannot be rebuilt from bytes on disk. The raw results are unavailable: the value is pending, not wrong.
- SCI-003, missing path (US). Both parties see the same output, but the path that produced it compares two hard-coded dictionaries and never loads an authoritative input. Agreement on the value cannot repair a path that does not exist. The structure is challenged.
- SCI-004, disagreement (UNRESOLVED). One mapping counts 24 manifested payload files, the other 25 files recursively. Neither is chosen. The outcome stays UNRESOLVED on the record.
- SCI-005, restated under a declared convention. An explicit, prospective counting convention (24 payload files plus 1 manifest control file) makes the mappings agree. It is a new determination; SCI-004’s UNRESOLVED outcome is not rewritten. Standing effect of the restatement: STRENGTHENED (PROPOSAL_REVISED).
Claim ceiling
- Promotion records that a result met this class's checks; it does not certify a discovery or establish causality.
- The worked examples map a review of software-governance experiments, not a natural-science study.
- Methodological inference in the source review is recorded as not adjudicated, and stays so here.
- A promoted claim has passed this class’s checks. It is not a certified discovery, a causal finding, or proof that every relevant mechanism was declared.
Class JSON · Class details and mutation battery · Try it in the inspector