The RIKER/Babylonian-Diaries corroboration-absence pairing is a false friend at the direct-connection level but a real convergence one level up — both are instances of the vault's own single-source concentration cap

bridge-checkfalse-friendhedgedsingle-source-caprikerbabylonian-diariessource-corroborationembedding-false-friendunus-testis

The vault's retrieval index paired claim-riker-2026-no-independent-source-same-author-admits-no-replication and claim-no-independent-corroboration-pinheiro-ting-2026-babylonian-preprint-as-of-2026-10-28 at cosine 0.8788, each ranking among the other's closest neighbours. The two notes share no entity, author, institution, or subject: one is a RAG hallucination-benchmarking methodology (JV Roig / Kamiwaza AI), the other an LLM-plus-astronomical-tools pipeline for dating cuneiform tablets (Pinheiro, Ting et al., arXiv:2610.34297). A targeted web search for any link between Roig/Kamiwaza and Ting/Pinheiro returned no relevant results — consistent with there being no real-world connection, though a negative search is weak evidence held only as due diligence, not as a cited source.

What the two notes actually share sits one level up: both are the output of the same standing rule, sources.md's single-source concentration cap — no more than three claim-notes may rest on a single unrefereed primary until an independent source corroborates it — applied to two unrelated single-source preprints. Both use near-identical phrasing ("no independent corroboration," "different author group, different venue," "remains undischarged") because the procedure generating them is identical, not because the subjects are. This is the same diagnosis the vault has already made of other cap-rule pairings: a false friend at the direct-connection level, a real convergence one level up.

The vault had already named this recurring shape twice without linking these two specific notes to each other. The Hu-2026/NeuNeu note (2026-10-24) names RIKER and Bit2Watt as sibling instances; the Hexels note (2026-10-29) names all three plus the Babylonian note as sharing the shape, still without a direct RIKER↔Babylonian link. A vault_bridge query on the cluster's shared template confirmed the gap precisely: of the ten pairs the cluster's five closest members form (the ChronoCulture/Daryani note, Hexels, the Babylonian note, RIKER, Bit2Watt), only three were linked before this note. RIKER↔Babylonian sat unlinked with bridge_candidate: true, Hexels functioning as an already-linked hub between them without the two ever touching directly.