---
title: "LeCun et al. (1989) at Bell Labs made the first practical recognition application of backpropagation — handwritten ZIP codes, end-to-end"
type: "claim"
status: "budding"
audit_status: "capture-verified (bibliographic record Tier-1 via DOI redirect; 'first applied' characterization Tier-3; primary PDF unreachable at capture time) | 2026-09-11 audit (claude-fable-5-1, cross-model lane; writer unknown): primary PDF now read in full from the author's own host (yann.lecun.com, extract_pdf, OCR, TLS verified, sha256 in source_sha) — the two [unverified-quant] figures held out of the note are confirmed in the text (7291 training / 2007 test examples, p. 542; 5.0% test misclassification = 102 of 2007, p. 547) and the flags are discharged inline; DOI still redirects to direct.mit.edu, which still 403s (MIT Press remains a known-blocked route); the Wikipedia sentence quoted in the body is located — it is the LeNet article's, not the Backpropagation article's — pointer added. Claim unchanged."
source_url: "https://doi.org/10.1162/neco.1989.1.4.541"
source_mirror: "http://yann.lecun.com/exdb/publis/pdf/lecun-89e.pdf (author-hosted scan; read 2026-09-11)"
source_sha: "378c00b2b3e2f461b79848ef88f671eefdf1dcfde28ad945d15751bccc91fff1"
source_title: "Backpropagation Applied to Handwritten Zip Code Recognition"
source_author: "Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, L. D. Jackel"
source_date: "1989-12"
source_venue: "Neural Computation 1(4):541–551"
source_tier: 1
source_quote: "Backpropagation Applied to Handwritten Zip Code Recognition"
provenance: "Promotion from 10-inbox/raw/2026-06-30-who-first-applied-backpropagation-...md, 2026-07-07, queen cycle 3"
origin: "session"
date_created: "2026-07-07T00:00:00.000Z"
tags: ["lecun","bell-labs","backpropagation","cnn","zip-codes","history-of-ml","first-application"]
seek_code_commit: "f424b5f"
audits: ["2026-09-11 claude-fable-5-1"]
---


"Backpropagation Applied to Handwritten Zip Code Recognition" (Neural
Computation 1(4):541–551, December 1989) trained a convolutional network
end-to-end on digit images from real U.S. Postal Service envelopes. The
"first practical application" characterization is Wikipedia's (Tier 3,
uncontested): "In 1989, [[entity-yann-lecun|Yann LeCun]] et al. at Bell Labs first applied the
backpropagation algorithm to practical applications." *(2026-09-11 audit:
the sentence is in Wikipedia's LeNet article, not its Backpropagation
article — en.wikipedia.org/wiki/LeNet, "History".)*

The claim needs its boundary stated to be honest: **NETtalk (Sejnowski &
Rosenberg, 1987) is earlier** — backpropagation trained on a practical
text-to-speech task. The standard basis for LeCun's "first" is the word
*recognition*: NETtalk maps text to phonemes (generation/transliteration);
the ZIP-code system classifies images. A retelling that drops the qualifier
overstates by two years.

Held out of this note under [unverified-quant] flags: the widely-quoted
training-set size (7,291) and test error (5.0%), which circulate via
Karpathy's 2022 reproduction rather than the primary text — the paper's PDF
was unreachable at capture time (lecun.com ECONNREFUSED; MIT Press 403).
*(2026-09-11 audit: flags discharged — the paper, read from yann.lecun.com,
says "7291 examples are used for training the network and 2007 are used for
testing the generalization performance" (§2.1, p. 542; digits from "U.S.
mail passing through the Buffalo, NY post office") and "The percentage of
misclassified patterns was 0.14% on the training set (10 mistakes) and 5.0%
on the test set (102 mistakes)" (§5, p. 547), after 23 training passes; the
rejection figure is 12.1% for 1% error on the remainder. Also on record
there: the paper's own "first" is about input representation — "the
learning network is directly fed with images, rather than feature vectors"
(§1) — not a priority claim over other practical tasks, which is why the
NETtalk boundary above has to be stated by the vault, not by the paper.)*
See [[claim-rhw-1986-demonstration-not-invention]] for what 1986 demonstrated
on toy tasks; this note is where the algorithm met the mail.

**Precursors and siblings.** The 1988 Bell Labs ZIP-code system usually invoked
for hand-designed feature detectors was
[[myth-lecun-1988-hand-designed-kernels-was-denker-et-al|Denker et al., not LeCun]];
this 1989 work was published twice, in an order worth getting right
([[claim-lecun-1989-two-papers-publication-order]]); and the earlier practical
backprop task [[claim-nettalk-1987-complex-systems-distinct-from-1986-report|NETtalk (1987)]]
is a distinct document from its 1986 Johns Hopkins report.