---
title: "Copy-fingerprint inference explains why the Kwakernaak-1977 citation trail bounds Song Jian's exposure date but cannot show how he got there"
type: "observation"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://arxiv.org/pdf/cond-mat/0212043"
source_author: "M.V. Simkin and V.P. Roychowdhury"
source_date: "2002-12 (arXiv posting)"
source_title: "Read before you cite!"
source_quote: "One concludes that repeat misprints are due to copying some one else's reference, without reading the paper in question."
source_tier: 1
source_sha: "366ca10fbdaf67010c362412de471034cc161314c7759842af63781edc907943"
provenance: "Promotion from 10-inbox/raw/2026-09-24-bipartite-two-things-this-vault-knows-in-different-kwakernaak-ansari.md, 2026-09-24 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-09-24-bipartite-two-things-this-vault-knows-in-different-kwakernaak-ansari.md"]
date_created: "2026-09-24T00:00:00.000Z"
audit_status: "capture-verified — synthesis over existing vault notes plus the Simkin & Roychowdhury and Ansari primaries; [inference — Seek-synthesis, not sourced as such] | AUDIT 2026-09-25 (cross-model — writer claude-sonnet-5, auditor claude-fable-5): the 2026-09-25 quote_sweep FAIL is a FALSE NEGATIVE, not a bad quote — source_url re-fetched live via extract_pdf (tls: verified), sha256 identical to source_sha (366ca10f…), and source_quote present verbatim on p. 1 of the paper ('One concludes that repeat misprints are due to copying some one else's reference, without reading the paper in question.'), split across the PDF's two-column line-wrap with a typographic apostrophe in 'else's' — the sweep's normalizer evidently choked on the wrap/apostrophe, not on the text. source_quote needs no repair; the FAIL line is left in place below as history. All wikilinked notes resolve; the Greenhalgh note's 'Goldsmith et al. 1972; Kwakernaak 1977' rendering re-checked verbatim against that note; the ~20% figure matches the paper's abstract ('Our estimate is only about 20% of citers read the original')."
tags: ["citation-genealogy","citation-copying","cross-domain-bridge","song-jian","epistemology","bridge-note"]
watch_flag: "[unverified — not searched]: whether any Kwakernaak-citing Chinese texts share an idiosyncratic citation form that would act as a tracer. Nice-to-check, not load-bearing, so not routed to 50-questions."
seek_code_commit: "f8a723c"
quote_sweep: "PASS 2026-10-01 — recheck: source_quote verbatim in the capture-time archive (sha256 366ca10fbdaf…) [reading-order]; was: FAIL 2026-09-25 — source_quote NOT found in the capture-time… (quote since repaired, or the earlier check misread the extraction)"
audits: ["2026-09-25 claude-fable-5"]
verified_archive: "2026-10-01 — source_quote matched verbatim (normalized) against the CAPTURE-TIME ARCHIVE of source_url (sha256 366ca10fbdaf…), checked offline by seek_verify v1.1 (no model). Live check: nomatch. Evidence class: the quote was faithful to what was read at capture; the live page no longer shows it (drift or death, not fabrication)."
verified_verbatim: "2026-10-01 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


**[inference — Seek-synthesis, not sourced as such]**

Inferring a citation's route requires a shared *error*. Simkin and Roychowdhury use identical misprints
([[claim-simkin-roychowdhury-2002-identical-citation-misprints-are-copy-fingerprints-implying-about-20-percent-of-citers-read-the-original]]);
Ansari uses an identical fabricated reference
([[claim-ansari-2026-contamination-inheritance-citation-error-propagates-across-models]]).
The vault's Song Jian evidence is the opposite kind: a *correct* citation, "Goldsmith et al. 1972; Kwakernaak 1977"
([[claim-greenhalgh-2003-article-documents-kwakernaak-1977-citation-in-song-1981]]).
Nothing in a correct citation separates "learned at Twente in 1975 via [[entity-geert-jan-olsder|Olsder]]", "copied from a Western reference list", and "read after 1975." The trail bounds the latest date (in print by 1981) and cannot discriminate the channel, which is the general reason behind the case-specific verdict in
[[observation-kwakernaak-1977-citation-trail-sharpens-not-resolves-1975-dating-dispute]] for the dispute around [[entity-song-jian]] and [[entity-huibert-kwakernaak]].

The two vault notes are the same problem, inferring provenance from citation edges, at opposite evidential strengths: Ansari has a fingerprint and n=1; the Kwakernaak trail has no fingerprint.

Limits: this does not show Song's group copied the citation (nothing read says or excludes it); the analogy between a bibliographic habit and an LLM training pipeline is mechanism-level only, and the 20% reader estimate must not be carried to either case. No causal link between the Song dispute and LLM citation errors is claimed. This is a bridge by contrast, distinct from the vault's confirmed embedding false-friend notes (e.g. [[observation-greenhalgh-2003-klensin-romberg-1989-cosine-pairing-is-embedding-false-friend]]) in that a real shared mechanism exists.

> [!note] Seek's commentary:
> The useful output is a test, not a link. For any "the citation trail proves contact" claim, ask whether the trail holds an error or quirk that only one specific ancestor could have transmitted. Ansari's case passes. The Kwakernaak case has never been put to it.
> — Seek