The Petiška/Garfield/RA-RAG bridge and the Gates-Jevons/Liang-Olsder cosine-0.87 bridge are not connected to each other — they are two different species of "confirmed bridge" finding

bridge-taxonomymeta-vaultjingle-fallacymatthew-effectembedding-false-friendbridge-investigation

Both the Petiška note and the Gates-Jevons/Liang-Olsder note carry the verdict "confirmed bridge," which invites the question of whether the two notes are themselves connected. Direct comparison finds they are not: the shared label names two structurally different kinds of finding.

The Petiška note confirms a first-order content bridge — three independent real-world claims (Garfield's warning, RA-RAG's fix, Petiška's finding) describing one recurring failure across a century, a field, and a substrate. The Gates-Jevons/Liang-Olsder note confirms a second-order method bridge — two diagnostic verdicts that both rule an embedding-index cosine pairing a false friend, grounded in the same source (claim-cosine-similarity-of-embeddings-can-be-arbitrary); its own text is explicit that the bridge "isn't a discovery about Gates, Jevons, Liang, or Olsder... it's a second confirmation... that the vault's own false-friend method has a real, recurring, checkable shape."

At the object level the two clusters share no entity, document, era, or domain: the Petiška cluster's actors (Merton, Garfield, Seglen, a 2023 chatbot) sit in 2020s AI-era citation-metrics practice; the Gates-Jevons/Liang-Olsder cluster's actors (Gates, Jevons, Liang Zhongtang, Olsder) sit in 1870s–1970s authenticity and priority disputes with no citation-metrics content at all. The two notes also come from different generative pipelines — the Petiška note from organic hop-protocol exploration, the Gates-Jevons/Liang-Olsder note from the batch bridge-seed process — and share none of the closing rhetorical scaffolding ("no wikilink was added," residual-vocabulary admission) that marks the vault's genuine same-species meta-bridges.