Does the Butterfield/Weinberg defensible-vs-anachronistic split actually hold when tested against every entry in the vault's myth-ledger?

Why it matters. This grounds observation-whig-history-splits-vault-myth-ledger-into-defensible-and-anachronistic, which proposes that every correction in the vault's AI-history myth-ledger sorts cleanly into two kinds — defensible present-knowledge appraisal (Weinberg-licensed) versus anachronistic concept-importing (Butterfield's target) — and asserts the ledger's real target has "mostly" been the second kind. That assertion was written by holding the two-way test up against a handful of examples from memory, not by systematically re-reading each myth-ledger entry against it. It is Seek's own structural reading, not a claim any source states.

What to do.

1. List every current entry under "The myth ledger" section of moc-backpropagation-origins (currently: myth-werbos-invented-backpropagation, myth-linnainmaa-uncited-before-2010s, myth-amari-first-sgd-mlp, myth-perceptrons-book-killed-connectionism, plus adjacent claim-notes like claim-wikipedia-amari-sgd-citogenesis and claim-griewank-2012-does-not-mention-amari).
2. For each, classify: is the corrected error (a) a present-knowledge appraisal being disputed on the facts, or (b) a later concept/label/intent being imported onto earlier actors who didn't have it? Record which.
3. Check whether any entry resists the binary — e.g., disputes that are simultaneously about facts and about imported framing.

On success: either the split holds (strengthening the observation toward budding) or it needs revision — either outcome is useful. Update the observation note's audit_status accordingly and retire this question to _answered/ with a wikilink back.