Citogenesis

The pattern where a claim originates in one (often hedged) source, gets repeated by secondary venues — Wikipedia foremost — without independent verification, and then reads as corroborated simply because it now appears in multiple places. The vault has run into this pattern repeatedly enough, across unrelated domains (AI-history priority disputes, disability-history myths, pseudohistorian labeling, and — as of this session — a search tool's confidently wrong "Hinton coined 'deep learning' in 2006" claim), that it earns a consolidated hub rather than living only inside individual claim threads. The companion failure mode is hedge erosion: a careful scholarly qualifier ("perhaps the first...") drops as the claim travels down-venue, which is a distinct mechanism from citogenesis proper (no one misquotes; the qualifier is just shed) but produces the same false-confidence result and the two are usually found together.

References

• claim-wikipedia-amari-sgd-citogenesis · myth-amari-first-sgd-mlp · claim-ivakhnenko-gmdh-first-deep-characterization · claim-chandler-wikipedia-labeled-pseudohistorian · claim-pignier-fabricated-braille-barbier-hostile-encounter-in-retirement-memoir · claim-hinton-2006-papers-omit-deep-learning-phrase
• Captures: 2026-07-03-did-robbins-monro, 2026-07-31-verify-the-origin-of-the-term-deep-learning