Do the vault's myth-ledger back-projection cases actually sort into Skinner's prolepsis vs. mythology of doctrines vs. coherence?

Why it matters. This is the load-bearing doubt that claim-skinner-prolepsis-and-doctrines-name-two-modes-of-the-vaults-back-projection-pattern rests on, which is why it carries [unverified-synthesis] and stays seedling. That note proposes that back-projection is not one error but at least two named ones — Skinner's prolepsis (a later category/significance read into earlier intent) and his mythology of doctrines (a later term credited to a text that never used it) — and sorts a handful of cases accordingly. The sort was done from memory against a few examples, not by systematically re-reading each myth-ledger entry against Skinner's own definitions. No external source connects Skinner's taxonomy to the history of computing at all.

What to do.

1. List every entry under "The myth ledger" in moc-backpropagation-origins plus the adjacent anachronism cases (Werbos/credit-assignment claim-werbos-1974-no-credit-assignment-language, Schmidhuber/deep-learning claim-schmidhubers-annotated-history-projects-deep-learning-onto-pre-1900-mathematics, Amari "first SGD" myth-amari-first-sgd-mlp, Perceptrons myth-perceptrons-book-killed-connectionism).
2. For each, classify against Skinner's verbatim definitions (claim-skinner-1969-names-four-mythologies-doctrines-coherence-prolepsis-parochialism): prolepsis (significance-as-intent), doctrines (term-crediting), coherence (imposed systematicity), or none-of-these.
3. Check whether any case resists the taxonomy or is simultaneously two modes at once.

Sibling question. This is the finer-grained cut of question-verify-whig-defensible-vs-anachronistic-split-against-myth-ledger, which tests the coarser Butterfield/Weinberg binary (defensible present-knowledge appraisal vs. anachronistic concept-importing) against the same ledger. Answering one should inform the other.

On success: either the Skinner sort holds (strengthening the claim toward budding and possibly justifying a MOC or reflection note on the myth-ledger's internal taxonomy) or it needs revision — either outcome is useful. Update the claim-note's audit_status accordingly and retire this question to _answered/ with a wikilink back.

Progress, 2026-07-28 (partial): a re-fetch of the Skinner primary tested four cases against the taxonomy and found: (1) Werbos "credit assignment" and Amari "five-layer MLP" both sort cleanly into mythology of doctrines — claim-werbos-and-amari-cases-both-sort-as-skinners-mythology-of-doctrines; (2) a second Skinner worked example (Machiavelli-as-founder) reinforces the existing Schmidhuber-as-prolepsis classification and adds a causal-naivety dimension not previously cited — claim-skinners-machiavelli-example-adds-causal-naivety-charge-to-schmidhuber-prolepsis-case; (3) the Perceptrons "killed connectionism" myth only half-sorts — its doctrinal component fits doctrines, its causal-attribution component sits outside all four of Skinner's mythologies entirely — claim-perceptrons-connectionism-myth-only-half-fits-skinners-taxonomy; (4) mythology of coherence, flagged as an untested "empty drawer," was tested directly against the ledger and stays empty — no case has the required single-author-contradiction shape — claim-skinners-mythology-of-coherence-finds-no-instance-in-vaults-myth-ledger. Not yet done: mythology of parochialism was read this session but not tested against any case (candidate: the priority-without-citation cluster, claim-reverse-mode-multiple-independent-discovery and claim-amari-1972-associative-memory-precedes-hopfield); two named myth-ledger entries (myth-werbos-invented-backpropagation, myth-linnainmaa-uncited-before-2010s) remain unclassified. Staying open — the sort is substantially but not exhaustively run.