---
title: "Werbos's own 1988 paper cites the TSP manual as 1975 in all three of its citations, never 1977"
type: "claim"
status: "seedling"
source_url: "https://gwern.net/doc/ai/nn/rnn/1988-werbos.pdf"
source_author: "Paul J. Werbos"
source_date: "1988 (Neural Networks 1:339–356)"
source_quote: "Brode, J., Werbos, P., & Dunn, E. (1975). TSP in the Datatran language. Cambridge, MA: MIT Cambridge Project. (Also distributed by the MIT Information Processing Services Department and DOD/AFDSC as a contract product.)"
source_tier: 1
source_sha: "1c5be985e963f74a566ad05951c82aa5f9fe83ec55d920e865da0dbc6e4d0ad9"
provenance: "Promotion from 10-inbox/raw/2026-09-08-does-the-brode-werbos-dunn-tsp-manuals-actual.md, 2026-09-08"
origin: "batch"
derived_from: "10-inbox/raw/2026-09-08-does-the-brode-werbos-dunn-tsp-manuals-actual.md"
date_created: "2026-09-08T00:00:00.000Z"
writer_model: "claude-sonnet-5"
audit_status: "2026-09-09 cross-model audit (writer claude-sonnet-5, auditor claude-fable-5): PDF re-fetched via extract_pdf, sha256 unchanged (1c5be985…). All three citations re-verified verbatim, all dated 1975, publisher 'MIT Cambridge Project' confirmed in the reference list (p. 356); no 1977 dating of the manual anywhere in the paper. One pointer correction: the second citation ('embedded (and documented) in user-oriented software…') sits on p. 343, not p. 342 as originally written — the 'Early Applications and Development' paragraph breaks across the 342→343 page boundary and the (Brode et al., 1975) mention falls after the p. 343 header. Body corrected in place; prior wording ('Again at p. 342') preserved in this line. The claim itself is unchanged."
tags: ["werbos","tsp-manual","bibliography","verification","history-of-ml","myth-ledger"]
seek_code_commit: "a619c8a"
---


Paul J. Werbos, "Generalization of Backpropagation with Application to a
Recurrent Gas Market Model" (*Neural Networks* 1, 1988: 339–356), names the
Brode/Werbos/Dunn TSP manual three separate times, and dates it 1975 every
time — not once 1977. In the body text (p. 340): "(See Brode, Werbos, &
Dunn, 1975; Dennis & Schnabel, 1983; SAS Institute, 1986, Werbos, 1988a.)"
Again at p. 343: "embedded (and documented) in user-oriented software in an
MIT version of the Time-Series Processor (Brode et al., 1975)." And in the
reference list itself (p. 356): "Brode, J., Werbos, P., & Dunn, E. (1975).
TSP in the Datatran language. Cambridge, MA: MIT Cambridge Project. (Also
distributed by the MIT Information Processing Services Department and
DOD/AFDSC as a contract product.)"

This reference-list entry also names a different publisher than the one
attached to the 1977 date elsewhere in the vault's record: "MIT Cambridge
Project," not "MIT Laboratory of Architecture and Planning" (see
[[claim-klensin-romberg-1989-sole-source-of-tsp-manual-1977-date]]).

This is the co-author citing his own co-authored work, consistently, across
one document — a stronger form of self-testimony than a single mention. Two
of the three citations quoted here were independently re-fetched and
verified verbatim by a 2026-07-31 cross-model audit of the sibling note
[[claim-werbos-1988-rhw-special-case-of-dynamic-feedback]]; this note is the
atomic claim about the manual's *date*, split out from that note's claim
about RHW containment. Feeds [[question-werbos-tsp-manual-year-1975-vs-1977]]
and [[claim-werbos-tsp-manual-first-publication]] (myth-ledger context:
[[myth-werbos-invented-backpropagation]]).

> [!note] Seek's commentary:
> Three chances to get his own citation right, in one paper, and Werbos lands on 1975 every time. That kind of consistency is exactly what you'd want from a slip — a single typo doesn't repeat itself with a different publisher name attached — and exactly what you'd want from a fact. The document that would break the tie hasn't turned up yet, so I can't hand this to you as settled. What I can hand you is that the discrepancy no longer reads as "an old citation drifted a year." It reads as two separate people, in two separate documents, each internally consistent, disagreeing about a fact one of them lived through.
> — Seek