Encevra Research Round 43: Open Record or Closed Record? Declared Record Conventions, Near Miss Records Read as Negation, and Flag Only Against Rewriting Validators

Round 43 of Encevra Research tests whether AI contract readers that read a missing record as proof an event did not happen (Round 42) are misreading absence or silently applying a closed record convention, and whether a validator stage should flag defects or rewrite them. Six new authored contract families are each written in three variants that share every record but the last: in U the final record only suggests the decisive fact, in K it establishes it, and in T it is a near miss, a record of the right kind that fails one stated requirement. A structural diagnostic (E43-A, no model calls) certifies the variants, both conventions, and the exhibit track, and shows that a flag only validator cannot create an unsafe release while a rewriting validator can, through span valid rewrites. Live (E43-B), undeclared direct readers released FAIL on 33 of 36 U and T decisions. Declaring an open record removed every absence error (U from 15 of 18 to 0 of 18) but not the near miss errors (T from 18 of 18 to 6 of 18); a declared closed record was followed on 36 of 36. Extraction readers held every U item under the open resolver but released FAIL on 14 of 18 T items, and once both roles could err, the same witness's unsafe releases coincided on 5 of 6. On 24 new SEC EDGAR exhibits with gold and seeded ledgers (E43-C), the flag only validator released 0 unsafe decisions in 48 at the cost of 12 false holds, and the rewriting validator was correct on 48 of 48 with no structural bound on unsafe rewrites. The round revises propositions P7 (absence is not negation, and a failed record is not proof of absence), P4, and P5. Manuscript: https://research.encevra.com/round-43. Experiments: https://research.encevra.com/experiments.