ifc-0238

17.4.1 How to read the evidence sequence

The chapter uses three nested scales of evidence. They answer different questions and should not be read as successive scores on one benchmark.

Scale	Scientific material	Question isolated	Strongest permitted conclusion
CTTE–0 through –5	Generated source ledgers with known hidden structure	Can a source-bearing presentation be reconciled, extended, queried, and transported?	Controlled recovery relative to a frozen document grammar
GLP1–PILOT–0/1	Human-curated GLP-1 trial events	Can a missing treatment context be represented or recovered without inventing biological claims?	Retrospective context construction in a bounded semantic interface
GLP1–GRANT	A frozen source ledger and typed prospective program	Can an admitted theory package be compiled into an auditable research program?	Structural and semantic realization results, not biomedical validity, peer-review quality, or fundability

The controlled CTTE worlds calibrate the machinery because their missing objects are evaluator-known. The GLP-1 pilots then replace generated claims by curated scientific records while retaining a narrow task. The grant sequence changes the output type from a retrospective theory packet to a prospective research program. Its final negative result is therefore part of the evidence: passing a formal prose proxy can still fail to demonstrate scientific or rhetorical quality.