What genuinely connects "No independent source corroborates RIKER (Roig, 2026)" and the Pinheiro/Ting Babylonian Diaries corroboration-absence note?

source-corroborationsingle-source-caphedged-bridgeembedding-false-friendbridge-seedrikerbabylonian-diariesvault-self-reference

Bridge-seed note

10-inbox/raw/ was searched first (by grep for both note slugs together) for an existing capture naming this exact pair. None exists — this is a fresh investigation, not a duplicate of a sibling session.

A direct mcp__seek__vault_pair check measured the pairing precisely: cosine 0.8788, each note sitting in the other's close neighbourhood (RIKER ranks 6th among the Babylonian note's neighbours; the Babylonian note ranks 4th among RIKER's), linked: false. A mcp__seek__vault_novelty check on this capture's own thesis ("both are instances of the vault's own single-source cap procedure, a hedged bridge not a real-world connection") returned max_cosine 0.776, novelty_percentile 71.2, verdict: adjacent — not duplicate-risk — and its five closest neighbours are other "shares genre, not substance" hedged-bridge observations already in the vault (the Pitman/Gabelsberger-IPA and Ossietzky/Liu-Xiaobo-Nobel pairing; the Polanyi/Ossietzky pairing; the Paul Allen person-bridge; the Puckette fourth-instance-hedged-bridge pairing), confirming this is a recognized vault genre and this specific instance of it is not yet recorded.

Claim: The two notes are not a substantive connection between RAG-benchmark hallucination research and Babylonian-astronomy dating — they are two independent applications of the same standing vault rule, the single-source concentration cap (sources.md, added 2026-07-29), to two single-source preprints that share no entity, author, institution, or subject matter

sources.md's own text states the rule both notes are executing: "No more than three claim-notes may rest on a single unrefereed primary — a preprint, a working paper, a self-published report, a conference abstract — until at least one independent source corroborates it. Independent means a different author group and a different venue." Both target notes are that rule's output, applied to unrelated subjects: RIKER (JV Roig / Kamiwaza AI) is a RAG hallucination-benchmarking methodology; the Pinheiro/Ting preprint (arXiv:2610.34297) is an LLM-plus-astronomical-tools pipeline for dating cuneiform tablets. Neither author, institution, or finding touches the other. A targeted web search for any direct link between JV Roig/Kamiwaza AI and Yuan-Sen Ting/Pinheiro returned no relevant results on either query — pure noise (an unrelated Japanese auto-filter company, a Yu-Gi-Oh! wiki page, a RIKEN webpage, and similar false matches) — consistent with there being no real-world connection to find, though a negative web search is weak evidence on its own and is not cited as a source here, only as due-diligence.

The resemblance the embedding index is picking up on is vocabulary, not referent: both notes are built from the vault's own templated procedure for this situation, and both use nearly identical phrasing because the procedure itself is identical — "no independent corroboration (different author group, different venue)," "single-source cap," "remains undischarged." This is the same shape the vault's own Greenhalgh-2003/Hu-2026 observation already names for a different pair: a false friend at the direct-connection level, a real convergence one level up. bridge_verdict: hedged.

Claim: This is not a one-off coincidence — it is a recognized, growing vault cluster, and two of its existing members already link both target notes to each other's neighbours without the two target notes ever having been linked directly to each other

The vault had already, independently of this capture, named this recurring shape at least twice. The Hu-2026/NeuNeu note (2026-10-24) states: "This is the same shape the vault has already recorded for other single-source-capped preprints: RIKER and Bit2Watt each show press, partnership, or venue activity around a preprint that is not the same as a second, independent research group testing it." The Hexels note (2026-10-29) goes further and names the exact pairing this capture was asked to investigate, without linking the two notes to each other: "This search shares its shape with the vault's other recent young-preprint corroboration searches — claim-hu-2026-no-independent-corroboration-found-as-of-2026-10-24, 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 — each a clean negative explained, at least in part, by the preprint's youth."

A direct mcp__seek__vault_bridge query on this cluster's shared template confirms the graph-level shape precisely: the top-5 hits for "no independent corroboration found for single-source preprint, single-source concentration cap remains undischarged" are the Daryani/ChronoCulture note, Hexels, the Babylonian/Pinheiro-Ting note, RIKER, and Bit2Watt — and among the ten pairs those five form, only three are already linked (Hexels↔Babylonian, Hexels↔RIKER, RIKER↔Bit2Watt). RIKER↔Babylonian — this capture's assigned pair — sits in the unlinked_pairs list, with bridge_candidate: true. Hexels functions as an already-linked hub between RIKER and the Babylonian note without those two ever having been connected to each other directly — exactly the gap this capture closes.

Claim: Even granting the shared procedure, the two notes' negative findings are not evidentially equivalent — RIKER's "no corroboration" is a seven-month-old absence confirmed in the author's own words, while the Babylonian note's is a three-day-old absence its own text attributes to simple recency

RIKER's note rests partly on the methodology's own creator stating, in a March 2026 follow-up paper (about seven months before this capture, and over five months after the original RIKER submission), "our findings have not been independently replicated with alternative frameworks" — a negative conceded by the single person with the most motivation to claim otherwise. The Babylonian note, by contrast, searched a preprint only three days past its 25 Oct 2026 submission and says so plainly: "Given the preprint's age at the time of the search, this absence reads as consistent with ordinary recency... rather than as a negative signal about the paper's reception." Both notes correctly apply the same cap rule and reach the same formal verdict ("cap remains undischarged"), but one is a mature, author-conceded absence and the other is an absence that has not yet had time to mean anything. Treating the two as equivalent instances of "unconfirmed" would overstate how much the Babylonian note's finding is currently worth.

Further leads

• The "no independent corroboration, single-source cap remains undischarged" template now has at least five confirmed members (RIKER, Bit2Watt, the FIC preprint, Hu-2026/NeuNeu, Hexels) plus this capture's Babylonian/Pinheiro-Ting note — six total — all generated since the cap rule's 2026-07-29 introduction. claim-gift-2026-unrefereed-and-unreplicated-as-of-2026-08-07 uses closely related language ("unrefereed and unreplicated") and was not checked this session for cluster membership. The cluster may now be large enough to warrant its own MOC on the model of moc-the-shape-recurs-the-object-does-not or moc-one-witness-is-not-enough (bees flag, the queen decides; not built here).
• mcp__seek__vault_bridge on the cluster's template returned bridge_candidate: true for more than just this capture's assigned pair — Bit2Watt is linked to RIKER but not to Hexels or the Babylonian note, which a future pass could check directly.
• Whether claim-no-independent-corroboration-pinheiro-ting-2026-babylonian-preprint-as-of-2026-10-28's verdict changes once the preprint is no longer "three days old" is an open, time-sensitive recheck — not pursued here since it would require a fresh corroboration search, not a bridge check.

Entity candidates

• Single-source concentration cap (sources.md, added 2026-07-29, Cali ruling) — concept — flagged first, per the known blind spot: this is the actual, older, foundational thing both target notes are instances of. The papers and their authors are incidental; the rule is the real ancestor the resemblance traces back to.
• bridge_verdict: hedged as a named vault pattern — concept — first articulated in observation-1981-testimony-convergence-and-werbos-gap-false-friend-share-lone-witness-floor (2026-10-05: "false friend at the mechanism level, a real connection one level up") and reused since for at least four other pairings; this capture is another instance of applying that pre-existing diagnostic, not a new one.
• JV Roig — person — RIKER's sole author (Kamiwaza AI); no further biographical detail chased this session.
• Yuan-Sen Ting — person — co-author of the Babylonian Diaries dating preprint (arXiv:2610.34297), self-hosts both the GitHub repo and project page per the existing claim-note.
• Pinheiro — person — lead/co-author of the same preprint per the vault's existing note title ("Pinheiro, Ting et al."); full given name not established in this vault as of this capture.

Safety flags

None. No PDFs or web pages were fetched this session requiring extract_pdf or archive_page — all substantive material came from already-sourced vault-internal notes (re-read directly, not re-fetched) plus two WebSearch due-diligence queries that returned only irrelevant noise (an unrelated Japanese auto-filter/mask company, a Yu-Gi-Oh! wiki page, a RIKEN institute page, and similar false matches on "RIKER"/"Ting"). No addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing encountered.