Catastrophic-forgetting and organizational-forgetting research independently converged on the same corrective move against the jingle fallacy

catastrophic-forgettingorganizational-forgettingjingle-fallacyfalse-bridgecross-domain-bridgeembedding-false-friendepistemics

Read together, claim-catastrophic-forgetting-is-a-structure-dependent-regime-not-universal and claim-david-brachet-2011-splits-forgetting-into-continuous-and-discrete-channels do not describe the same mechanism, and neither literature knows the other exists. At the level of mechanism, the resemblance the vault's embedding search flagged (cosine 0.86, unlinked) is a false friend, joining observation-ml-ad-en-gedi-cosine-pairing-is-embedding-false-friend and observation-hawks-jeffress-cosine-pairing-is-embedding-false-friend as roughly the vault's fourth logged instance of a cosine pairing that shares rhetorical shape rather than causal substance.

But one level up, both fields independently made the same kind of correction to a naive prior. A term — "forgetting" — that looked like it named one phenomenon turns out, on inspection, to split into a catastrophic/discontinuous regime and a graceful/gradual regime, gated by a structural condition specific to the domain: representational task-overlap for neural networks, stay-versus-exit for organizations. This is the general pattern the jingle fallacy names — treating a shared label as evidence of a shared construct, described under that name by Edward L. Thorndike in 1904 and paired with its mirror, the jangle fallacy, by Truman Lee Kelley in 1927 — run in reverse: rather than a fallacy committed and left uncorrected, both the neural-network literature and the David & Brachet decomposition are cases of a field catching and resolving exactly that error about its own central term, without either field being aware the other had done the same thing.