17.1 The transition from simulation to literature

The two testbeds address adjacent parts of a scientific loop:

Chapter 16 asks whether a structural proposal survives action in an executable world. This chapter asks whether heterogeneous reports can be converted into a precise proposal worth testing there. In a complete discovery cycle, document-grounded construction proposes a mechanism and a counter-witness experiment; a simulator, laboratory, or clinical study then tests it; the resulting evidence returns to the source-bearing theory store.

The chapter has two goals. The first is representational: preserve who claimed what, in which context, with which evidence, without collapsing disagreement. The second is constructive: determine whether a small extension of the current theory derives new, locked consequences and identifies a decisive external test. The second goal depends on the first, but must not be confused with causal identification.

The Theory Extension Package introduced below is this chapter’s realization of the artifact dossier from Section 10.4.1. It binds the proposed theory map to source-level provenance, transported evidence, locked corpus tests, and an executable external-admission interface. A corpus can support the package and reject alternatives; it cannot fill the external admission field merely by describing the experiment that would be needed. The name is reserved for proposals that actually change the presentation or generative language. A revised observer, retrieval policy, evidence workflow, or admission interface without such a map is recorded as a package revision, not promoted to theory extension by terminology alone.

The CTTE ladder progressively removes supplied entity, context, query, and domain structure while retaining source provenance and conservative abstention.