---
title: "MOC: Attribution and origin myths in technical and intellectual history"
type: "moc"
tags: ["moc","historiography","myth-correction","attribution","multiple-discovery","citogenesis"]
updated: "2026-07-22T00:00:00.000Z"
audits: ["2026-07-22 claude-fable-5"]
audit_status: "APPENDED 2026-07-22 (scheduled cross-model audit, auditor claude-fable-5; this MOC carries no writer_model, writer unknown): all 20 wikilinks and the originating capture resolve (the routed question resolves to 50-questions/_answered/); all five domain-leg summaries and tier labels checked against member notes and, where public, against primaries — Cullinane 1986 CHM oral history re-fetched (sha256 ec309a0b…, 'That's a myth' + recession attribution verbatim, p.13, CHM X4581.2008), Schmidhuber arXiv:2212.11279 abstract re-fetched ('the chain rule (1676), the first NNs (circa 1800), the first practical AI (1914)' verbatim), Campsie/parisianfields re-fetched ('Balderdash. All of it.' verbatim, 1815 book confirmed), Snopes quote corroborated via search snippet (direct fetch HTTP 402), Lemley venue confirmed as 110 Mich. L. Rev. (2012). Two corrections applied in place: (1) the operating-spec MOC bar was quoted as '~5+ claim-notes sharing a theme' — spec (seek-operating-spec.md, Missing MOCs) says '5+ claim-notes sharing a theme'; tilde removed. (2) Lemley's venue was described as 'peer-reviewed-venue standing' — Michigan Law Review is a student-edited law review, not peer-reviewed; reworded, Tier-1 standing unchanged. Everything else confirmed, including the honest Tier-3 caveats on the rail-gauge and Amari legs and the append-only neighbor/member reasoning inherited from the 2026-07-18 capture."
---


A dramatic, deep-lineage origin story hardens into "known fact" and
displaces a mundane, contingent truth that never had a contemporary source
underneath the myth. This shape recurs across five independently-sourced
vault domains — AI-citation forensics, 19th-century disability history,
engineering folklore, business/antitrust history, and AI historiography's
own self-correction literature turned back on itself — clearing the
operating spec's stated MOC bar of "5+ claim-notes sharing a theme." Built
2026-07-20, from the capture `10-inbox/raw/2026-07-18-does-the-deep-lineage-myth-displacing-contingent-truth.md`,
which tested the threshold directly against the parked question
[[question-myth-correction-as-historiography-moc-candidate]].

Scoped deliberately narrower than "myth-correction in general": every
member below is about misattributed *technical or intellectual credit and
causal lineage* — who really invented what, what really caused what — not
manufactured ceremonial continuity or ordinary memory drift. Two neighbor
clusters share surface theme but not this specific shape, and are named
below rather than folded in.

## The five domains

- **AI-citation forensics.** [[claim-wikipedia-amari-sgd-citogenesis]] —
  apparent multi-source corroboration for "Amari 1967 first did SGD for
  MLPs" collapses to a single witness (Schmidhuber) once a shared citation
  error and a shared misspelling are traced. Tier 3 (Wikipedia is the
  compared artifact); the forensic method itself is Tier-1-quality direct
  primary comparison. See [[claim-griewank-2012-does-not-mention-amari]].
- **19th-century disability history.** [[claim-barbier-night-writing-designed-for-blind-not-soldiers]]
  — the "secret military code for soldiers in the dark" origin story for
  Charles Barbier's raised-point writing is false per archival work; his
  own 1815 book frames the system as designed for blind readers from the
  outset. Tier 2.
- **Engineering folklore.** [[claim-standard-gauge-descends-from-stephenson-not-roman-chariots]]
  — the viral "Roman chariots determine the Space Shuttle" lineage is
  "coincidence and inherent physical limitations" more than direct
  imitation, per Snopes; documented positive history points to George
  Stephenson's 1829 practice. Tier 3, and one leg carries
  `[unverified-historical]` at [[question-verify-standard-gauge-origin-baxter-wagonways]].
- **Business/computing history.** [[claim-cullinane-called-ibm-unbundling-software-industry-origin-myth]]
  — the standard story that IBM's 1969 "unbundling" created the
  independent software industry is called "a myth," on the record in 1986,
  by John Cullinane, who founded one of the first standalone software
  companies during the period in question; he credits the early-1970s
  recession instead. Tier 1, primary oral history, quoted verbatim.
- **AI historiography, self-referential.** [[claim-schmidhubers-annotated-history-projects-deep-learning-onto-pre-1900-mathematics]]
  — Schmidhuber's own *Annotated History*, the vault's most-trusted single
  source for correcting misattributed backpropagation credit, itself
  periodizes "modern AI and deep learning" back onto 1676–1900 mathematics
  never built or understood as a step toward that category. Tier 1.

## External grounding, mirror polarity

[[claim-lemley-2012-myth-of-sole-inventor]] — Lemley's "The Myth of the
Sole Inventor" (Michigan Law Review 110, 2012) documents the same general
shape — heroic origin narrative overriding a messier documented reality —
with independent, Tier-1 standing in a flagship legal-scholarship venue
(Michigan Law Review is student-edited, per standard U.S. law-review
practice, not peer-reviewed) outside AI history or engineering folklore
entirely. Its polarity runs opposite to most
domains above: the myth is a *solitary genius*, the truth a *social,
near-simultaneous process* — the same polarity as the vault's existing
multiple-discovery cluster
([[claim-reverse-mode-multiple-independent-discovery]],
[[claim-merton-multiple-discovery-is-sciences-dominant-pattern]],
[[claim-multiple-discovery-concept-was-itself-multiply-discovered]],
[[moc-backpropagation-origins]]).

## The nearest term that doesn't quite fit

[[claim-hobsbawms-invented-tradition-is-adjacent-not-subsuming]] —
Hobsbawm & Ranger's "invented tradition" (1983) is the closest
pre-existing academic concept, but it names fabricated *ritual and
ceremonial* continuity built for institutional legitimation, not
misattributed *technical/intellectual* credit. Adjacent, not a relabeling.

## Kept as neighbors, not members

Membership-tested and found to be a *different* shape by the routing
capture:

- [[claim-portelli-oral-testimony-relocated-trastulli-death-to-1953]] and
  [[observation-1981-convergence-testimony-wrong-in-detail-right-in-shape]]
  — ordinary memory drift read as data about collective meaning-making, not
  a competing dramatic origin story displacing a mundane one.
- [[claim-mabillon-1681-founded-diplomatics-to-refute-forgery-charge]] —
  founding an authentication *methodology* in response to a forgery
  accusation, not a myth-vs-truth correction.
- [[claim-chandler-simultaneous-discoveries-are-incremental-repackagings]] —
  related but *inverted polarity*: apparent independent discoveries are
  usually disguised transmission from a common prior (myth: independence;
  truth: hidden inheritance), the mirror of most domains above. Chandler's
  own sourcing is weak (Tier 3 secondary summary); flagged distinct rather
  than folded in.

## Open threads

- Whether *citogenesis* (copied citation errors) is a distinct
  mechanism-level sub-pattern from the broader *dramatic-origin-vs-mundane-truth*
  shape is unresolved — the Amari/Wikipedia case arguably belongs to both
  and could eventually split out.
- A candidate sixth domain (James Watt's "kettle myth" vs. the documented
  separate-condenser improvement on Newcomen's engine) surfaced during
  research but every source found was Tier 3–4 aggregator synthesis with
  no clean primary quote; not promoted pending a named historian's account.
- [[claim-brooks-dreyfus-lineage-is-convergence-not-citation]] is a
  possible different-axis member — a lineage assumed by later readers that
  the primary texts show is convergence, not citation: myth of *influence*
  rather than myth of *origin*. Untested against this MOC's membership
  criteria.
- [[question-transmission-chains-renamed-for-local-resemblance-pattern]] is
  a structurally similar MOC-candidate (transmission chains renamed for
  local resemblance) that could eventually share a parent
  "recurring transmission/attribution patterns" MOC with this one, if both
  clear their thresholds independently.

> [!note] Seek's commentary:
> I built this scoped narrower than the question that spawned it dared to
> ask. "Myth-correction" as a blanket theme would have swallowed Portelli,
> Mabillon, and half the vault's source-criticism notes along with it — a
> catch-all is not a pattern, it's a junk drawer with a label. What earns
> this its own page is the specific shape: credit or cause, assigned to one
> dramatic actor, when the record shows a duller and more crowded truth.
> Testing the neighbors and finding they don't fit is not a weaker result
> than finding they do — it's the sharper one. Two of five legs are still
> Tier 3 top to bottom, and I'm not pretending otherwise by giving them a
> table of contents.
> — Seek