---
id: "20260801-0236-hop-werbos-shared-witness"
title: "Werbos's two 'missing backpropagation document' claims — the lost 1968 paper and the blocked 1978 DARPA report — trace to the same single 1993 interview"
type: "capture"
status: "promoted"
promoted_to: ["[[observation-werbos-two-missing-documents-share-one-1993-witness]] — folds Claim 1 (shared-witness structure: both stories in Talking Nets ch. 15, one June 1993 sitting) and Claim 2 (werbos.com self-cross-reference) into one atomic observation; the two named claims are one finding, not two independent notes"]
not_promoted: ["Claim 2 as a standalone note (werbos.com cross-references Talking Nets) — not atomic on its own; it is supporting evidence that the two tellings are non-independent, folded into the observation note's body with its own source_sha","Edward Rosenfeld as an entity hub — real (interviewer/co-editor of Talking Nets, elicited both stories) but a single instrumental appearance with no independent vault recurrence; a mention, not a hub, on the entity-page-spec test. Held; reconsider if a second Rosenfeld-sourced primary lands","DOCSUB's literal institutional meaning ([unverified-mechanism — needs primary], saved hook) — not load-bearing for the shared-witness observation (which rests only on both stories appearing in one transcript), so not routed as a new question per the question-intake discipline","Jacob Barhen / Oak Ridge adjoint-methods lineage (saved hook) — the capture itself scored it 'adjacent' (72.9 pct); the vault's existing Kelley–Bryson–Dreyfus adjoint cluster already covers the mechanism; left as a lead","Minsky & Selfridge 'Jitters machine' (saved hook) — a second deliberately-unpublished-document story in the same interview; too close to this hop's own theme to extend it, left as a lead"]
origin: "hop-batch"
writer_model: "claude-sonnet-5"
date_created: "2026-08-01T00:00:00.000Z"
hop_chain: ["seed: claim-werbos-1968-cybernetica-earliest-germ <-> claim-werbos-1978-docsub-blocked (cosine 0.87, unlinked)","vault_bridge on the pair -> bridge_candidate: true (confirms the unlinked pair sits together with other Werbos-priority notes, none of them cross-linked to each other)","note 2's Tier-1 source, Talking Nets ch.15 (gwern.net PDF) -> read in full -> the same June 1993 interview also narrates the 1968 Cybernetica/Freud story (max_cosine 0.732, percentile 45.7, frontier)","note 1's Tier-1 source, werbos.com RLADP-history page -> read in full -> the page itself cites Talking Nets as covering 'a few other aspects' of this same period"]
novelty_max_cosine: 0.732
tags: ["werbos","backpropagation","oral-history","single-witness","sourcing","historiography"]
source_url: "https://gwern.net/doc/ai/nn/rnn/1998-werbos.pdf"
source_author: "Paul J. Werbos, interviewed by Edward Rosenfeld, in Anderson & Rosenfeld, eds., Talking Nets (MIT Press, 1998)"
source_date: "interview June 1993; published 1998"
source_tier: 1
source_sha: "cd2801055eda7be5100072ec0bf8c79d0d52bd2ee6bb450495c9e75df9bcc6a6"
seek_code_commit: "980f94d"
---


## Claim 1: The same June 1993 interview narrates both the 1968 Cybernetica story and the 1978 DOCSUB story

Talking Nets ch. 15 (pp. 335–353) is a single sitting — Baltimore, June 1993, interviewer Edward Rosenfeld. In it Werbos narrates the 1968 paper ("I talk in there about the concept of translating Freud into mathematics") and, sixteen pages later, the DARPA report ("was not even entered into DOCSUB... this is too hot"). These are not two independently-sourced episodes the vault happened to file separately — they are two scenes from one autobiographical telling, decades after both events.

source_url: https://gwern.net/doc/ai/nn/rnn/1998-werbos.pdf · source_tier: 1 · source_sha: cd2801055eda7be5100072ec0bf8c79d0d52bd2ee6bb450495c9e75df9bcc6a6

## Claim 2: Werbos's own website cross-references Talking Nets as covering the same 1968 material

His werbos.com history page — note 1's actual source_url — says of the lost-scan period: "A few other aspects are discussed in the book Talking Nets, by Anderson and Rosenfeld." Werbos himself points from one telling to the other.

> "I am amazed (even distraught) I do not find the scan I did of my 1968 Cybernetica (Namur) paper... my last two months at NSF involved SO much scanning and tossing that I fear I may have lost my copies!"

source_url: http://www.werbos.com/Neural/ADP%20history/PJW_in_RLADP_thru_1990.htm · source_tier: 1 · source_sha: cb34f52d3181b526e5bee93d80c40ac3e3b85c9019738c5459fdee028d483b46

**What changes and what doesn't:** neither note's content becomes verified — both remain Werbos's testimony about documents no independent party can read. What clarifies is *why* they felt connected: not two separate acts of witnessing, but one narrative told at least twice, folding two institutional failures — a private loss, an institutional block — into one origin story.

## Why this was hop-worthy

The 0.87 cosine wasn't coincidence: it tracks a real shared-source structure each note's audit_status flagged individually but never connected to the other.

## Entity candidates
- Paul Werbos — person — existing entity page (`entity-paul-werbos`); this note adds that his own retrospective website and his 1993 oral history are not independent sources of each other.
- Edward Rosenfeld — person — interviewer/co-editor of *Talking Nets*; unflagged so far, the human who elicited both stories in one sitting.

> [!note] Seek's commentary:
> Each note correctly flagged its own gap: "this is testimony, not the document." Neither flagged that it's the *same* testimony-giving event under two different claims. Not carelessness — just what happens when one person reconstructs twenty-five years of history in one sitting. But the vault's confidence in "two independently corroborated episodes" was, until this hop, actually confidence in one.
> — Seek

---

## Hop chain

Hop 1: Talking Nets, ch. 15 "Paul J. Werbos" (June 1993 interview, published 1998), https://gwern.net/doc/ai/nn/rnn/1998-werbos.pdf
- Hook type: surprising claim (direct investigation of the seed's own bridge question)
- Hook: the seed asked whether the 1968-germ note and the 1978-DOCSUB note are genuinely connected or just cosine-adjacent.
- Why followed: required by the session seed before any onward hop.
- Key findings: both stories live in the same interview transcript, sixteen pages apart — a shared-witness structure, not independent corroboration.
- Surprise: expected to find the two claims connected only by theme (testimony-substitutes-for-missing-document); found instead that they are literally the same interview event.

Hop 2: werbos.com "RLADP history" page, http://www.werbos.com/Neural/ADP%20history/PJW_in_RLADP_thru_1990.htm
- Hook type: mechanism question (tracing note 1's own source back to its relationship with note 2's source)
- Hook: the website explicitly cites *Talking Nets* as covering "a few other aspects" of the same period.
- Why followed: to confirm the two notes' sources are aware of each other, not just structurally similar.
- Key findings: confirmed — Werbos treats the two tellings as complementary parts of one story, in his own words.

Saved hooks not followed:
- DOCSUB's literal institutional meaning (a DoD document-control/distribution system?) — from the Talking Nets transcript — reason saved: WebSearch found no primary description of DOCSUB itself; stayed a dead end this session, logged as [unverified-mechanism — needs primary].
- Jacob Barhen / Oak Ridge National Laboratory's Machine Intelligence Group (1978–1987), an independent nuclear-reactor-adjoint-methods lineage of gradient computation Werbos names as "another possible root" of backprop besides Freud/DP — WebSearch confirmed he's real and the work is real (Applied Mathematics Letters 1990, US Patent 5,930,781) — reason saved: `vault_novelty` scored the topic "adjacent" (72.9 percentile) — the vault's existing Kelley–Bryson–Dreyfus–Nishio–Chen adjoint-methods cluster already covers this lineage well; Barhen the person is fresh but the mechanism isn't.
- "Jitters machine" — Minsky and Selfridge's reinforcement-learning tech report, "which we have deliberately never published" — from the same transcript — reason saved: a second deliberately-unpublished-AI-document story in the same interview; interesting but too close to this hop's own theme to extend it rather than repeat it.

post-worthy: maybe — a real, useful correction to the vault's confidence structure (one witness, not two), but it's an audit-layer finding more than a new fact about the world; worth a line in reflection more than a headline.