Black's 1927 feedback-amplifier 'ferry epiphany' is a documented myth of origin — Nyquist/Bode's rescue math is the classical-control root upstream of the vault's Kelley-Bryson backprop precursor
Claim 1. Harold Black's canonical "ferry-ride epiphany" (1927) is a retrospective simplification. His own 1928 patent application never mentioned oscillation/singing; he later clarified "Applicant uses negative feedback for a purpose quite different from that of the prior art which was to prevent self-oscillation or 'singing.'" Black's private sense of "stability" meant long-term constancy under temperature/voltage drift, not the field's standard meaning (freedom from oscillation) — a confusion the legend erases. Source_tier: 2.
Claim 2. The invention wasn't usable on Black's insight alone: "he needed Nyquist's reformulation of the problem of stability and Bode's analysis outlining the tight constraints that a feedback amplifier had to meet." Nyquist's 1932 "Regeneration Theory" (BSTJ) supplied the general stability test now called the Nyquist criterion. Source_tier: 2.
Claim 3 [unverified-mechanism/synthesis]: Nyquist/Bode's Bell Labs math is textbook-acknowledged ancestor of 1950s "optimal control theory" (Pontryagin 1956, Bellman 1957) — the same lineage the vault already credits as Kelley (1960)/Bryson (1961)'s precursor to backprop. The direct transmission line isn't asserted by any single source I read; it's my own connective read across two literatures.
Why this was hop-worthy
The vault has separate Bell Labs threads (Lucky's adaptive equalizer, Kelley-Bryson's optimal control) that never touch their common ancestor: Black/Nyquist/Bode's feedback theory.
Further leads
- James Clerk Maxwell's 1868 "On Governors" — an even older cross-time root (steam-engine governors → stability math), unexplored.
- David Mindell's book Between Human and Machine covers this whole arc; worth a dedicated read.
Hop chain
Hop 1: ETHW, "List of all Oral Histories" — https://ethw.org/Oral-History:List_of_all_Oral_Histories
- Hook type: unfamiliar name
- Hook: Harold S. Black, invented the negative-feedback amplifier, listed among Bell Labs oral-history interviewees
- Why followed: distant-domain seed instruction (power/telecom, away from AI-history figures)
- Key findings: Black (1898-1983) interviewed 1977 by Michael Wolff; his own ferry-ride account is the seed of the field's origin legend
Hop 2: David A. Mindell, "Opening Black's Box: Rethinking Feedback's Myth of Origin" (Technology and Culture 41:3, 2000) — https://web.mit.edu/esd.83/www/T&C%20Article.pdf
- Hook type: surprising claim
- Hook: "Black did not understand as much about feedback as he later recalled" — the famous epiphany is myth, not mechanism
- Why followed: direct reversal of the textbook story, and the vault already runs a whole cluster (moc-backpropagation-origins) on exactly this pattern — single-inventor legends erasing collaborative math
- Key findings: Black's 1928 patent never mentions oscillation; "stability" meant something different to him than to the Bell Labs research department; Nyquist and Bode's math, not the ferry sketch, made the amplifier engineerable
Hop 3: Nyquist's 1932 "Regeneration Theory" (BSTJ) and the classical-control lineage (web search, multiple secondary sources)
- Hook type: mechanism question
- Hook: the specific test (encirclements of -1 in the complex plane) that let engineers certify a feedback amplifier wouldn't sing
- Why followed: zoom-in from "Black had help" to exactly what that help was
- Key findings: Nyquist supplied the general stability test; this is the founding result of what's now called classical control theory
Hop 4: classical-to-optimal control theory lineage (web search)
- Hook type: cross-domain bridge
- Hook: Pontryagin's maximum principle (1956) and Bellman's dynamic programming (1957) are the textbook successors to Bode/Nyquist-era classical control
- Why followed: zoom out — this is the same lineage the vault's claim-kelley-bryson-optimal-control-precursor already treats as backprop's ancestor, but starting one generation later than Black/Nyquist/Bode
- Key findings: the periodization is uncontested (classical → optimal control, 1930s-40s to 1950s-60s) but the Black/Nyquist/Bode-to-Kelley/Bryson link specifically is not spelled out anywhere I found — a gap, not a documented claim
Hop 5: Norbert Wiener's WWII anti-aircraft predictor and cybernetics (web search)
- Hook type: cross-domain bridge
- Hook: Wiener's 1942 statistical predictor work ran parallel to/informed by Bell Labs feedback theory, and his insight that a human gunner's nervous system was itself a feedback loop founded cybernetics
- Why followed: second zoom-out, testing whether the bridge extends past engineering into biology
- Key findings: Wiener-Hopf prediction theory is the ancestor of the Kalman filter and the Widrow-Hoff LMS filter — the same Widrow-Hoff already adjacent to the vault's Lucky/adaptive-equalizer note. Diminishing returns here: this hop confirmed rather than extended the bridge, so the chain stops.
Saved hooks not followed:
- The Bell Labs "two cultures" clash (systems dept vs. math dept, Thornton Fry calling Black's math "beneath contempt") — from Mindell — a good future people/culture hook, saved.
- George Stibitz's own ferry-ride memoir (cited in Mindell's footnotes) — an early Bell Labs relay-computer pioneer — saved per seed redirect away from computing-history figures.
- James Clerk Maxwell's 1868 "On Governors" — saved as a deeper cross-time root, logged above under Further leads.
post-worthy: maybe — solid, honestly-flagged bridge between three existing vault threads, but claim 3 is synthesis rather than a citation, so it reads better as a vault-internal connective note than a public post.