---
id: "20260728-0200-do-the-vaults-myth"
title: "Do the vault's myth-ledger back-projection cases actually sort into Skinner's prolepsis vs. mythology of doctrines vs. coherence?"
type: "capture"
status: "promoted"
promoted_to: ["30-notes/claim-werbos-and-amari-cases-both-sort-as-skinners-mythology-of-doctrines.md","30-notes/claim-skinners-machiavelli-example-adds-causal-naivety-charge-to-schmidhuber-prolepsis-case.md","30-notes/claim-perceptrons-connectionism-myth-only-half-fits-skinners-taxonomy.md","30-notes/claim-skinners-mythology-of-coherence-finds-no-instance-in-vaults-myth-ledger.md","40-entities/entity-thomas-kuhn.md","40-entities/entity-arthur-c-danto.md"]
not_promoted: ["Mythology of parochialism vs. the priority-without-citation cluster (Reverse-mode multiple discovery, Amari 1972/Hopfield) — read this session but not tested against any case; left as a lead in the existing open question rather than a new claim, since nothing was actually concluded.","WebSearch finding no scholarship extending Skinner's taxonomy to science/technology history — a further-leads pointer, not a claim; its substance is already carried by the [unverified-synthesis] flag on all four promoted notes, so a separate note would duplicate the flag rather than add information.","Danto/Winiarski primaries still unread — a reading lead, not a claim; folded into the new Danto entity page instead of a standalone note.","Entity candidate E. H. Gombrich — real but thin: a single footnote-chain mention (Skinner's other cited root for 'paradigm' vocabulary), no independent connection to the vault's domain beyond that footnote. Left as a mention, not promoted to a hub, per the entity spec's bias against the flood.","Entity candidate Mark Bevir — a single unread secondary-literature lead (a 2011 Intellectual History Review piece on Skinner's 'myth of coherence'), not yet load-bearing for any kept claim. Left as a mention, not promoted.","Entity candidate 'mythology of parochialism' (concept) — named and defined in the vault already via the four-mythologies claim-note; not promoted to its own entity page this session because it has not yet been tested against a single vault case, so a hub would front-run the actual finding."]
origin: "batch"
writer_model: "claude-sonnet-5"
date_created: "2026-07-28T00:00:00.000Z"
provenance: "batch run, 2026-07-28"
derived_from: []
tags: ["historiography","quentin-skinner","prolepsis","mythology-of-doctrines","mythology-of-coherence","myth-ledger","backpropagation-origins","anachronism","synthesis","verification"]
source_urls: ["https://dl.icdst.org/pdfs/files/38605e2c189ce8b26d3dc348bb359b1d.pdf"]
source_authors: ["Quentin Skinner"]
source_dates: ["1969 (essay); PDF re-fetched 2026-07-28"]
source_tiers: [1]
---


**Short answer the claims below support:** Partially, and unevenly across Skinner's four mythologies. This capture re-fetched Skinner's 1969 primary directly (`extract_pdf`, TLS verified, not a re-read of the vault's cached quotes) and tested it against four already-catalogued myth-ledger cases rather than the two the vault had previously checked. Result: the narrowly *terminological* cases (a specific word or technical framework retroactively credited to a text that never contained it) sort cleanly into **mythology of doctrines** — this now holds for two independent cases, not one. The **prolepsis** classification of the Schmidhuber "deep learning" case is reinforced by a second, closer-matching worked example inside Skinner's own text that the vault hadn't previously cited. But **mythology of coherence** — the drawer flagged as "standing empty" in [[claim-skinner-1969-names-four-mythologies-doctrines-coherence-prolepsis-parochialism]]'s commentary — stays empty on this closer look: no myth-ledger case tested here is a true instance, because coherence-mythology is specifically about imposing false unity on *one thinker's own body of work*, and the vault's field-level narrative cases don't have that shape. And one case — the Perceptrons "killed connectionism" myth — only half-sorts: it fuses a piece that fits Skinner's doctrines mythology with a piece (multi-decade, multi-actor causal attribution) that Skinner's taxonomy, built for textual-interpretive claims about individual thinkers, does not address at all. So the vault's cases sort into Skinner's taxonomy only insofar as they are claims *about what a specific text says or means* — they resist it wherever they are instead claims about a field's periodization or causal trajectory.

---

## Claim: Two independent myth-ledger cases — Werbos's "credit assignment" (1974) and Amari's "five-layer MLP" (1968) — sort cleanly into Skinner's mythology of doctrines

**Claim type:** Synthesis / classification. This is a definitional application of an already Tier-1-verified taxonomy to material already established elsewhere in the vault at Tier 1; the mapping itself is this capture's own comparative reading, so it carries `[unverified-synthesis]` even though every fact it rests on is independently sourced.

Skinner's mythology of doctrines names the error of "converting some scattered or quite incidental remarks by a classic theorist into his 'doctrine'" — attaching a vocabulary, framework, or systematic position to a text that never articulated it. Two vault cases fit this shape precisely, not by analogy but by the same mechanism (a specific *term* or *technical framing* is applied to a document that contains neither):

- **Werbos/"credit assignment."** [[claim-werbos-1974-no-credit-assignment-language]] establishes (Tier 1, direct grep over the cached 453-page extraction) that the words "credit" and "assignment" occur zero times in Werbos's 1974 thesis. The retrospective label "backprop solves spatial credit assignment (Werbos 1974)" credits the thesis with a doctrine — a named technical achievement — it never states in its own words.
- **Amari/"five-layer MLP with two modifiable layers."** [[myth-amari-first-sgd-mlp]] and [[claim-amari-1968-saito-experiment-primary-read]] establish (Tier 1, direct OCR read of Amari's 1968 book, pp. 94–135) that the words 層/多層 (layer/multilayer) are absent from the entire scanned span describing Saito's experiment; the model described is a piecewise-linear discriminant, not an MLP in the book's own terms. "Five-layer MLP" is Schmidhuber's later re-description, not Amari's.

Both are the same move Skinner catalogues under doctrines: a later commentator supplies a name/framework the original text did not use and then treats the text as though it had. Two independent instances sorting the same way strengthens the case that "doctrines" (not "back-projection" generically) is the right label for the vault's term-crediting cluster specifically, distinct from the category-projection cluster below.

| Field | Value |
|---|---|
| source_url | https://dl.icdst.org/pdfs/files/38605e2c189ce8b26d3dc348bb359b1d.pdf |
| source_author | Quentin Skinner |
| source_date | 1969 |
| source_tier | 1 |
| exact_quote | "This mythology takes several forms. First, there is the danger of converting some scattered or quite incidental remarks by a classic theorist into his 'doctrine' on one of the mandatory themes." |
| supporting_vault_claims | [[claim-werbos-1974-no-credit-assignment-language]] (Tier 1); [[myth-amari-first-sgd-mlp]] and [[claim-amari-1968-saito-experiment-primary-read]] (Tier 1) |

---

## Claim: Skinner's own worked example of prolepsis — Machiavelli as "founder of the modern political orientation" — structurally mirrors the vault's Schmidhuber "1676 chain rule opens deep learning" case, reinforcing the existing prolepsis classification with a source the vault hadn't previously cited

**Claim type:** Synthesis / classification, reinforcing an existing vault claim rather than duplicating it. [[claim-schmidhubers-annotated-history-projects-deep-learning-onto-pre-1900-mathematics]] and [[claim-skinner-prolepsis-and-doctrines-name-two-modes-of-the-vaults-back-projection-pattern]] already classify Schmidhuber's periodization as prolepsis using Skinner's Petrarch example; this claim adds a second, structurally closer worked example from the same primary that the vault had not cited.

Immediately after the Petrarch/Renaissance example, Skinner works a second case that is a tighter structural match for the vault's Schmidhuber problem than Petrarch is: Machiavelli being called "the founder of the modern political orientation." Skinner grants this may be a true statement of *retrospective significance* while insisting it is not, and cannot be, a true statement of Machiavelli's own *intended meaning*: "this may well provide a true account of Machiavelli's historical significance (though it seems to presuppose a somewhat naive view of historical causation)." That parenthetical is the closer match to the vault's case: Skinner is not merely rejecting the founder-label as anachronistic (the Petrarch point) — he is separately flagging that treating a person as "founder of" a later-named field also smuggles in an unearned *causal* claim. Schmidhuber's Annotated History does exactly this twice over: it reads "deep learning" back onto the 1676 chain rule (the significance/meaning conflation, already classified as prolepsis) and implicitly casts that mathematics as the causal origin of a field it could not have been intended to found (the same "naive view of historical causation" Skinner flags in the Machiavelli case, parenthetically, without developing it further).

| Field | Value |
|---|---|
| source_url | https://dl.icdst.org/pdfs/files/38605e2c189ce8b26d3dc348bb359b1d.pdf |
| source_author | Quentin Skinner |
| source_date | 1969 |
| source_tier | 1 |
| exact_quote | "Machiavelli, we are very often told, 'was the founder of the modern political orientation.' ... Now this may well provide a true account of Machiavelli's historical significance (though it seems to presuppose a somewhat naive view of historical causation)." |
| footnote_context | Skinner cites Warren Winiarski, "Niccolo Machiavelli" in *History of Political Philosophy*, ed. Strauss and Cropsey, 247, for the "founder" quotation (his fn. 100). |
| supporting_vault_claims | [[claim-schmidhubers-annotated-history-projects-deep-learning-onto-pre-1900-mathematics]]; [[claim-skinner-prolepsis-and-doctrines-name-two-modes-of-the-vaults-back-projection-pattern]] |

---

## Claim: The "Perceptrons book killed connectionism" myth only partially sorts under Skinner's taxonomy — a doctrinal component fits mythology of doctrines, but its field-level causal-attribution component falls outside all four of Skinner's named mythologies

**Claim type:** Synthesis / classification. `[unverified-synthesis]` — this is the capture's own decomposition of a myth already documented in the vault, not an externally sourced equivalence.

[[myth-perceptrons-book-killed-connectionism]] records that Minsky & Papert's 1969 *Perceptrons* is popularly credited with having "proved neural networks were a dead end and single-handedly caused the field's ~15-year stall." That myth splits cleanly into two components under Skinner's definitions, and only one of them is a Skinner mythology:

- **The doctrinal component fits mythology of doctrines.** [[claim-perceptrons-multilayer-sterile-was-conjecture]] establishes that the book's multilayer-pessimism line was a self-flagged "intuitive judgment" with an explicit invitation to reject it, not a proof. Compressing a hedged, incidental remark into a firm "the book proved X" doctrine is exactly Skinner's first form of the doctrines mythology — "converting some scattered or quite incidental remarks by a classic theorist into his 'doctrine.'"
- **The causal component has no home in Skinner's taxonomy.** The claim that the book "caused the field's ~15-year stall" is not a claim about what Minsky and Papert meant or said (which is the entire domain Skinner's essay addresses); it is a claim about multi-actor, multi-decade institutional and funding history — the same terrain [[myth-perceptrons-book-killed-connectionism]] shows is genuinely multi-causal (the Mansfield Amendment, the Lighthill Report, the absence of any hidden-layer training method until 1986). Skinner's four mythologies are built entirely around interpreting a single past writer's intended meaning or the coherence of a single writer's own corpus; nothing in the 1969 essay addresses causal explanation of a field's trajectory across many actors and institutions. This is not a case that "resists" the taxonomy through ambiguity — it is a case that is genuinely half inside Skinner's scope and half outside it, because the myth itself fuses two different kinds of claim that Skinner never fuses.

This directly answers the routing question's item 3 ("check whether any case ... is simultaneously two modes at once"): the answer here is not "two of Skinner's modes at once" but "one Skinner mode plus one non-Skinner mode, fused into a single popular myth."

| Field | Value |
|---|---|
| supporting_vault_claims | [[myth-perceptrons-book-killed-connectionism]]; [[claim-perceptrons-multilayer-sterile-was-conjecture]] (both Tier 3/1 as recorded in their own notes) |
| reasoning_basis | Skinner 1969 primary, full read (this capture) — the essay's stated scope is confined to "what are the appropriate procedures to adopt in the attempt to arrive at an understanding of the work" of "a work of literature ... or ... a work of philosophy"; no passage addresses field-level causal history. |

---

## Claim: Mythology of coherence has no clean instance among the vault's tested myth-ledger cases — the drawer flagged as empty stays empty on this closer look

**Claim type:** Synthesis / classification (a negative finding). `[unverified-synthesis]`, but the negative is now based on a direct full read of the coherence passage rather than an unexamined gap.

Skinner's mythology of coherence is specifically about a historian's temptation to "supply or find in each of these texts the coherence which they may appear to lack" — imposing a unified system on *one thinker's own body of work*, discounting that thinker's own explicit statements or entire sub-bodies of work where they would spoil the unity (his worked examples: Hobbes's "religious whole," Locke-as-lifelong-liberal despite his authoritarian thirties, Marx-and-Engels treated as inseparable to preserve "one corpus"). None of the myth-ledger cases tested in this capture or the prior session have that shape. The closest structural cousin is Schmidhuber's Annotated History, which does impose a continuous narrative line ("the chain rule (1676), the first NNs (circa 1800), the first practical AI (1914)") across otherwise disconnected episodes — but this is a modern historian narrating *many different authors across centuries*, not a historian imposing false unity on *one author's own contradictory output*. That distinction is load-bearing for Skinner's definition (his examples turn on discounting one author's own stated intentions or one author's own inconvenient texts), so the Schmidhuber case remains classified as prolepsis, not coherence. [[claim-skinner-1969-names-four-mythologies-doctrines-coherence-prolepsis-parochialism]]'s commentary flagged coherence and parochialism as "drawers standing empty" that hadn't been tested against a single vault entry; this capture tests coherence directly and finds it still empty, not because no vault case resembles it, but because the vault's back-projection cases are mostly about multi-author field narratives, a genre Skinner's coherence-mythology was never built to describe.

| Field | Value |
|---|---|
| source_url | https://dl.icdst.org/pdfs/files/38605e2c189ce8b26d3dc348bb359b1d.pdf |
| source_author | Quentin Skinner |
| source_date | 1969 |
| source_tier | 1 |
| exact_quote | "it will become dangerously easy for the historian to conceive it as his task to supply or find in each of these texts the coherence which they may appear to lack." |
| exact_quote_2 | "This procedure gives the thoughts of various classic writers a coherence, and an air generally of a closed system, which they may never have attained or even been meant to attain." |

> [!note] Seek's commentary:
> The shape of the answer surprised me more than either "yes" or "no" would have. I expected the sort to fail on some case being ambiguous — two mythologies both plausible, argue it either way. Instead it failed on scope: Skinner built this taxonomy to police one very specific move (an interpreter versus one dead author's own words), and the vault's myth-ledger is mostly doing something one level up — narrating a *field's* history across dozens of authors and institutions. The cases that survive the transplant cleanly (Werbos, Amari) are the ones that happen to still be about one text's own vocabulary. The cases that don't (Schmidhuber's periodization, the Perceptrons causal myth) are field-level from the start, and no amount of careful reading makes Skinner's drawers the right size for them. That's not a failure of the mapping — it's a more precise finding than "it mostly works": it works exactly as far as the vault's cases are actually about individual texts, and it should be expected to keep failing wherever they aren't. — Seek

---

## Further leads

- Skinner's mythology of parochialism (misattributing "influence" between writers from superficial similarity, with his own three-part sufficiency test: genuine similarity; the doctrine unavailable elsewhere; low probability of independent coincidence) was read in full this session but not tested against any vault case — a strong candidate check against the vault's priority-without-citation cluster ([[claim-reverse-mode-multiple-independent-discovery]], [[claim-amari-1972-associative-memory-precedes-hopfield]]).
- The full myth-ledger list named in [[question-classify-myth-ledger-by-skinner-prolepsis-vs-doctrines]] is not exhausted: [[myth-werbos-invented-backpropagation]] and [[myth-linnainmaa-uncited-before-2010s]] remain unclassified against Skinner's taxonomy.
- A WebSearch this session for scholarship applying Skinner's "mythology of doctrines/coherence" specifically to the history of science or technology (rather than intellectual/political history, his own domain) returned only intellectual-history and philosophy-of-education discussions (e.g. Mark Bevir, "Skinner and the 'Myth of Coherence,'" *Intellectual History Review* 21:1, 2011 — unread, Tier 2 lead) — no source found extending the taxonomy to science/technology history. The application to the vault's AI-history material remains the vault's own extension, not an externally attested one. `[unsourced — needs verification if load-bearing]`
- Arthur C. Danto, *Analytical Philosophy of History* (Cambridge, 1965), p. 169 — Skinner's own cited source (fn. 96) for the "concepts only available at a later time" argument underlying prolepsis; still unread in the vault after two sessions citing it secondhand.

## Safety flags

No safety flags encountered this session. The Skinner 1969 primary (icdst.org PDF mirror of the JSTOR-digitized scan) was re-fetched via `extract_pdf`; its provenance recorded `tls: "verified"`, so the elevated-suspicion rule for weak transport does not apply. No addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing was found anywhere in the 52-page extraction. The WebSearch results (academic summaries, a Haverford course blog, Studocu/PhilPapers listings) were read only for pointers, not treated as sources for any claim above the sourcing floor.

## Entity candidates

- Thomas Kuhn — person — Skinner explicitly credits *The Structure of Scientific Revolutions* (1962) for "the priority of paradigms," the mechanism underlying all four of Skinner's mythologies ("this notion of the priority of paradigms has already been very fruitfully explored... in the history of science" — Skinner's fn. 18). This is the foundational figure Skinner's own method is built on, not a co-author of the piece being classified — flagged first per the standing blind-spot note.
- E. H. Gombrich — person — Skinner's other cited root for "paradigms" vocabulary, from *Art and Illusion* (1960), one step further back than Kuhn in the same footnote chain.
- Arthur C. Danto — person — see Further leads; the philosophical ground under prolepsis specifically ("concepts only available at a later time"), cited but still unread by the vault after two sessions.
- Mark Bevir — person — author of a 2011 *Intellectual History Review* piece specifically on "Skinner and the 'Myth of Coherence'"; a secondary-literature lead for deepening the coherence classification, unread this session.
- mythology of parochialism — concept — the fourth named mythology, read in full this session (misattributed "influence," misapplied familiar categories) but not yet tested against any vault case; candidate for its own claim-note once tested.