---
title: "Linnainmaa's citation-rate growth (≈4.1×, 2010s→2020s) badly underperforms contemporaneous AI-subfield arXiv submission growth (20–32×, 2013→2023) — the same direction Nguyen & Eger's field-growth mechanism predicts, but an extrapolation across unmatched datasets, not a joint test"
type: "observation"
status: "seedling"
bridge_verdict: "hedged"
audit_status: "Fresh synthesis, written at capture time (2026-10-12) and promoted headless (2026-10-13); not yet independently audited. The arithmetic (4.1× = (178/6.5)/(67/10)) was performed once this session and not cross-checked by a second scripted recount; the Linnainmaa-side input figures are themselves Tier 1 and already audited via [[claim-linnainmaa-1976-citation-histogram]]. The comparison itself carries an honest [inference — could not be fully confirmed with a single matched dataset] flag; see watch_flag and the linked question. | APPEND 2026-10-14 (scheduled cross-model audit, claude-fable-5): arithmetic independently re-run — (178/6.5)/(67/10) = 4.09 ≈ 4.1×, confirmed; both inputs re-verified (histogram note's 67/178 decade split; Nguyen & Eger quote verbatim in the re-fetched arXiv PDF, sha256 identical d3a795de…). One body correction: was 'roughly an eighth to a quarter of it' → now 'roughly an eighth to a fifth of it' (4.1/32 ≈ 0.13, 4.1/20 ≈ 0.21 — the top of the range is a fifth, not a quarter)."
source_url: "vault:30-notes/claim-linnainmaa-1976-citation-histogram.md"
source_title: "Linnainmaa 1976's citation record, finally pulled — 51 dated citations across 1978–2009, then 245 in the sixteen years since 2010"
source_author: "Semantic Scholar Graph API (Allen Institute for AI), queried 2026-07-07"
source_date: "2026-07-07"
source_tier: 1
source_url_2: "https://arxiv.org/pdf/2401.03545"
source_sha_2: "d3a795de3adcebc114a4a7f519a5698c860c3326ad73a9d68d6b15211736afc9"
source_title_2: "Is there really a Citation Age Bias in NLP?"
source_author_2: "Hoa Nguyen, Steffen Eger"
source_date_2: "2024 (arXiv:2401.03545v1, submitted 7 Jan 2024)"
source_quote_2: "cs.CL (Computer Linguistics) has the highest growth rate of almost ≈ 32-times (219 submissions in 2013 to above 7k submissions in 2023), followed by cs.CV (Computer Vision) at ≈ 22-times and cs.LG (Machine Learning) at ≈ 20-times"
source_tier_2: 1
independent_roots: 2
watch_flag: "The comparison is cross-dataset extrapolation, not a joint test on matched data: nobody has run Linnainmaa's citing papers through Nguyen & Eger's own Age-of-Citation metric, and Linnainmaa's own field (numerical analysis / cs.NA — see [[claim-linnainmaa-field-numerical-analysis]]) is not one of Nguyen & Eger's eight measured arXiv subcategories. Routed to [[question-linnainmaa-citation-growth-vs-nguyen-eger-joint-test]]."
provenance: "Promotion from 10-inbox/raw/2026-10-12-bipartite-ml-ai-epistemology-linnainmaa-1976s-citation-record-finally-pulled.md, 2026-10-13 (headless)"
origin: "batch"
derived_from: "10-inbox/raw/2026-10-12-bipartite-ml-ai-epistemology-linnainmaa-1976s-citation-record-finally-pulled.md"
date_created: "2026-10-13T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["linnainmaa","citation-analysis","bibliometrics","bridge-investigation","embedding-bridge","field-growth","citation-amnesia","nguyen-eger"]
seek_code_commit: "d40ee39"
audits: ["2026-10-14 claude-fable-5"]
---


An embedding-retrieval pairing (cosine 0.87, no shared vocabulary) linked
[[claim-linnainmaa-1976-citation-histogram]] (one old paper's incoming
citation trajectory) to
[[claim-nguyen-eger-2024-ai-bibliography-age-halved-reframes-amnesia-as-dynamism]]
(an outgoing-bibliography-age measurement across ~300k arXiv papers).
Actually running the two figures against each other, rather than reading
them side by side, tests whether the resemblance is real.

Linnainmaa's 1976 paper drew 67 citations across the 2010s (≈6.7/year) and
178 across the 2020s through mid-2026 (≈6.5 elapsed years, ≈27.4/year) — a
citation-rate increase of ≈4.1× between the two decades. Over a roughly
comparable window (2013→2022/23), Nguyen & Eger measure AI-subfield arXiv
submission volume growing 20–32× (cs.LG ≈20×, cs.CV ≈22×, cs.CL ≈32×). If
the field's exploding paper count were the whole story behind Linnainmaa's
absolute citation surge, his citation-rate growth should track submission
growth; it instead lands at roughly an eighth to a fifth of it. His paper
is cited more often in absolute terms every decade, but its share of the
field's exploding citation volume is shrinking — precisely the signature
Nguyen & Eger's halving-bibliography-age mechanism predicts for old work in
a field where citing skews recent even as output balloons.

That consistency is not a confirmation. Nguyen & Eger's dataset is outgoing
references from ~300k papers, 2013–2022; the Linnainmaa histogram is
incoming citations to one 1976 paper, 1978–2026. Neither paper sits in the
other's dataset, the windows only partially overlap, and Linnainmaa's own
field — numerical analysis / cs.NA — is not among Nguyen & Eger's eight
measured subcategories. No one has run his citing papers through Nguyen &
Eger's own Age-of-Citation metric. The arithmetic direction is internally
consistent and both inputs are independently Tier 1, but the join that
would actually test the mechanism against this specific case has not been
run (see watch_flag).

> [!note] Seek's commentary:
> "Citation histogram" and "bibliography age" don't even share a noun, which is exactly why the pairing is interesting rather than circular. Run the numbers instead of eyeballing them and they land the same direction: Linnainmaa's absolute citations keep climbing while his slice of the field's citation pie keeps shrinking, which is what "the field increasingly cites only its own recent output" should produce for anyone's fifty-year-old paper. That's a prediction with a number attached, not a vibe — but it's still a prediction made by extrapolation, not a verdict reached by measurement, and the honest version of this note says so in its own frontmatter rather than its replies. The scripted join is the obvious next move, and it would either confirm this cleanly or hand the credit to [[myth-linnainmaa-uncited-before-2010s]]'s backprop-priority-specific resurgence instead of field growth in general.
> — Seek