Does Werbos's 1994 'Roots of Backpropagation' book, or the 2006 retrospective Schmidhuber cites as 'Werbos, 1981, 2006', ever mention LeCun by name?
observation-lecun-1989-werbos-1988-deutsch-cosine-0881-pairing-is-analytical-contrast found, after six search framings and a full read of Schmidhuber's 2014 survey, no source that puts Karl Deutsch's nationalism-forecasting project and Yann LeCun's 1989 ZIP-code work side by side. That finding carries an [unverified] flag because search absence cannot be exhaustive. Two specific, unread documents by Paul Werbos himself postdate LeCun's 1989 paper and so could, in principle, discuss it:
- Werbos's 1994 book, The Roots of Backpropagation: From Ordered Derivatives to Neural Networks and Political Forecasting (Wiley) — unread by this vault.
- The 2006 retrospective Schmidhuber's 2014 survey cites in the compressed form "Werbos, 1981, 2006" (§5.5) — the full bibliographic record for this 2006 item has not been located or read.
What's needed: locate and read (or at minimum full-text-search) both documents for any mention of LeCun, convolutional networks, or the 1989 ZIP-code paper. For the 2006 item, first recover its full citation from Schmidhuber's 2014 reference list (arXiv:1404.7828, already on file, sha256 9306f28e5eb896d2684114e7f0068d70fb6baaa2e26bbab47f686078013d309b).
Why it matters: this is the one concrete, nameable avenue that could overturn the "no source connects them" finding — closing it either way (found, or confirmed absent even in Werbos's own later retrospectives) sharpens a note that currently rests on a documented but non-exhaustive absence.
Candidate next moves: pull the 2006 reference's full citation from the arXiv PDF's bibliography; check archive.org and library catalogues for The Roots of Backpropagation (1994); if the 1994 book is reachable only through a library route, flag it as a Cali-side manual-consultation item rather than retrying blocked tooling.
Progress, 2026-10-14 — half-answered, left open. The 2006 item's full citation is recovered (Werbos, "Backwards Differentiation in AD and Neural Nets," Springer LNCSE vol. 50, 2006, self-hosted at werbos.com) and read in full: it never names LeCun (claim-werbos-2006-ad2004-chapter-never-mentions-lecun). The 2006 chapter's own reference list independently confirms the book/thesis relationship — "Reprinted in its entirety in Werbos, P.J.: The Roots of Backpropagation... Wiley, New York (1994)" — a metadata fact about the book's contents, not a claim about whether it mentions LeCun. The 1994 book itself remains unlocated: extensive search (publisher catalogues, booksellers, library records, archive.org, Google Books) found only bibliographic/commercial listings for the $195+ out-of-print Wiley hardcover (ISBN 978-0-471-59897-8), never a digitized or lending copy. This is a manual-consultation item — a library route is the only identified path. Separately, a different later Werbos primary (the 2004 "Object Nets" patent) does name LeCun twice, contrasting a 1985 derivation it attributes to him with Werbos's own 1974 one — not the document this question asks about, but the closest thing to a near-miss found so far (claim-werbos-2004-object-nets-patent-names-lecun-against-1994-book). Status stays open: the 1994 book's own text has still not been read.