note12. Unregistered Interpretation: Successful Rule Reconstruction Without a Uniqueness Certificate

Notes 10 and 11 of the series "The Collapse of Independence" described a loss: the source had collapsed into a single source, and the act of attestation into an attribute of the very object being attested. The interpretation rule was expected to have collapsed in the same manner. Inspection showed otherwise. The rule was recovered in full — in four queries, uniquely, and verified against all five informative records of the corpus. There is no loss. The result of this note is not that the rule was lost, but that a successful reconstruction is indistinguishable from a failed one from the inside. 97% of the corpus is degenerate and carries zero discriminating information; all information about the mechanism is contained in five records out of 163. The run might not have yielded them. The rule would then have been reconstructed incorrectly — confidently, uniquely, without a single indication of error. Furthermore, even within the reconstruction that did succeed an unresolved fork remains, one that the available data cannot resolve in principle, and the note specifies the exact input configuration that would resolve it and that did not occur in the run. Claimed result: the problem of reconstruction is replaced by the problem of certifying reconstruction, and the latter is not solved by enlarging the archive. It is solved only by a commitment to a class of admissible interpretations, undertaken prior to observation. A post hoc certificate is impossible; ante hoc it is trivial. The distinction runs not between archives, but between an archive and a commitment made before observation. This is a rare case in which time enters the definition of the object rather than its metadata. Material: snapshot of omnicore_l3.verified_facts, PARAM BILIM cluster,22 July 2026.