---
title: "Hahn and Bentley (2003) apply Kimura and Crow's neutral-allele model to baby-name popularity and conclude a value-neutral process with no preference or selection reproduces its power-law distribution"
type: "claim"
status: "seedling"
source_url: "https://doi.org/10.1098/rsbl.2003.0045"
source_author: "Matthew W. Hahn, R. Alexander Bentley"
source_date: "2003"
source_title: "Drift as a mechanism for cultural change: an example from baby names"
source_venue: "Proc. R. Soc. Lond. B (Suppl.) 270, S120-S123 (2003), DOI 10.1098/rsbl.2003.0045 — read via a Berkeley course-hosted mirror PDF (https://www.stat.berkeley.edu/~aldous/Research/Ugrad/Hahn-Bentley.pdf); the Royal Society DOI page is the venue of record and was not independently fetched this session"
source_quote: "In population genetics, it has been shown analytically (Kimura & Crow 1964) that the number of neutral alleles"
source_quote_2: "The distribution of baby names in the United States and their turnover in the twentieth century can be explained by"
source_tier: 1
source_sha: "4d101514313d1da1e9ed6963880f10725e407b9b35a5b3c8be34a6d0811d5ed7"
source_url_2: "https://www.stat.berkeley.edu/~aldous/Research/Ugrad/Hahn-Bentley.pdf"
source_sha_2: "4d101514313d1da1e9ed6963880f10725e407b9b35a5b3c8be34a6d0811d5ed7"
source_quote_3: "This simple model is equivalent to the infinite-allele model of population genetics for a single-locus, multiple neutral-allele system (Kimura & Crow 1964; Crow & Kimura 1970)."
source_url_3: "https://www.stat.berkeley.edu/~aldous/Research/Ugrad/Hahn-Bentley.pdf"
source_sha_3: "4d101514313d1da1e9ed6963880f10725e407b9b35a5b3c8be34a6d0811d5ed7"
source_quote_4: "a completely value-neutral process: no preference, fitness, or selection is needed, only proportional sampling."
source_url_4: "https://www.stat.berkeley.edu/~aldous/Research/Ugrad/Hahn-Bentley.pdf"
source_sha_4: "4d101514313d1da1e9ed6963880f10725e407b9b35a5b3c8be34a6d0811d5ed7"
independent_roots: 1
provenance: "Promotion from 10-inbox/raw/2026-09-28-hop-kimura-crow-neutral-drift-bridges-citation-copying-and-llm-dilution.md, 2026-10-10 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-09-28-hop-kimura-crow-neutral-drift-bridges-citation-copying-and-llm-dilution.md"]
date_created: "2026-10-10T00:00:00.000Z"
writer_model: "claude-opus-4-8"
audit_status: "capture-verified — the source capture (writer claude-sonnet-5) records both fragments as grounded via quote_check against the fetched mirror PDF (source_sha 4d101514…); this promotion ran headless with no network, so no independent re-fetch was performed. The read of record is the mirror PDF, whose fidelity to the Royal Society edition is an assumption until a second copy or the print edition agrees. Awaiting the verifier bee's model-free verbatim sweep against the archived copy. | CORRECTED 2026-10-11 (cross-model audit, claude-opus-5): the mirror PDF was independently re-fetched via extract_pdf (tls verified) and returned the same sha256 4d101514…, read end-to-end. Both recorded quotes are verbatim. Two corrections. (1) Title said the paper applies Kimura and Crow's 'formula'; the paper states the analytic Kimura–Crow result and then declines to fit it — 'Unfortunately, the data available to us do not represent an instantaneous distribution, as they are the cumulative number of baby names in a decade. We therefore modelled the random copying of names by computer simulation' — so what is applied is the equivalent infinite-allele *model*, by simulation; title and body now say so, and the whole sentence carrying that equivalence is added as source_quote_3. (2) source_quote_2 was a mid-sentence clause and is now the whole sentence it was cut from. source_quote is left cut before the inline formula on purpose: pdftotext renders 'ux−1(1−x)u−1' as 'ux 21 (1 2 x)u 21', so the remainder of that sentence cannot be copied as the source prints it. | APPEND 2026-10-13 (Cali's daily review, Claude Fable 5.1): receipt repair, proven against the capture-time archive (sha 4d101514…): source_quote_2 crosses a column/page-furniture break in the mirror PDF (the running header sits between \"explained by\" and \"a completely\"), so the whole string fails; split into source_quote_2 + source_quote_4, no words changed (match: reading-order, reading-order) → pointer 00-meta/cali-checklist.md §2"
watch_flag: "Source was read via a Berkeley course-hosted mirror PDF, not the Royal Society DOI page (which was not independently fetched this session). Per sources.md, the mirror is an allowed source with the venue of record named; its fidelity to the published Proc. R. Soc. B (Suppl.) edition is not independently confirmed. | 2026-10-11 (auditor, claude-opus-5): mirror-fidelity risk downgraded but not closed — the re-fetched file is the Royal Society's own typeset offprint, carrying a 'Downloaded from rspb.royalsocietypublishing.org on July 15, 2010' stamp on every page, the journal's running heads, and its printed pagination S120–S123, which is what a publisher PDF looks like rather than a retyped copy. The DOI page itself remains unfetched."
tags: ["cultural-evolution","neutral-theory","population-genetics","baby-names","power-law","drift","cross-domain-bridge"]
seek_code_commit: "f8a723c"
quote_sweep: "PASS 2026-10-11 — source_quote verbatim (normalized) in the capture-time archive (sha256 4d101514313d…), same-night model-free sweep | PASS 2026-10-11 (auditor re-fetch, claude-opus-5) — all three source_quotes read verbatim (line-wrap normalized) in a fresh extract_pdf of the same mirror URL, sha256 unchanged"
audits: ["2026-10-11 claude-opus-5"]
---


In "Drift as a mechanism for cultural change: an example from baby names," Matthew W. Hahn and R. Alexander Bentley carry the same 1964 population-genetics result the bibliometrics literature borrowed into a wholly unrelated popularity phenomenon. They cite it directly — "In population genetics, it has been shown analytically (Kimura & Crow 1964) that the number of neutral alleles" with a given frequency follows a predictable distribution — and argue that the power-law distribution of U.S. first-name frequencies over roughly a century is reproduced by "a completely value-neutral process: no preference, fitness, or selection is needed, only proportional sampling." Names, in this account, rise and fall in popularity not because parents prefer them but because each generation copies names in proportion to how common they already are; the resulting turnover and concentration fall out of drift alone. The paper frames itself explicitly as a null model against selection-based explanations of cultural change.

What is borrowed is the *model*, not a curve-fit of the 1964 formula. Hahn and Bentley quote the Kimura–Crow analytic result and then set it aside — their Social Security Administration data are "the cumulative number of baby names in a decade" rather than an instantaneous distribution, so they "therefore modelled the random copying of names by computer simulation," a simulation they describe as "equivalent to the infinite-allele model of population genetics for a single-locus, multiple neutral-allele system (Kimura & Crow 1964; Crow & Kimura 1970)." The shared object is the neutral-allele model and its parameter u = 4Nm; the published formula is cited as the analytic statement of it, not used as the estimator.

This is the middle leg of the lineage set out in [[observation-kimura-crow-1964-is-shared-named-ancestor-of-citation-copying-and-cultural-drift-models|the lineage observation]]: the identical [[entity-neutral-drift-model|neutral-drift]] formula that [[claim-simkin-roychowdhury-credit-kimura-crow-1964-neutral-allele-model-as-prior-derivation|Simkin and Roychowdhury credit for their citation-count model]] also underwrites a cultural-evolution model of name popularity, reached in a different field and without borrowing from the bibliometrics side — this paper (2003) predates Simkin's preprint (2005) and cites Kimura & Crow on its own account. What the 2026-10-11 audit removed is the stronger reading: Simkin's own text names this paper as an example of the same theory applied to culture, so the two uses are parallel but not mutually unaware. It is a second, distinct cultural-evolution mechanism imported into the vault alongside the iterated-learning/bottleneck account in [[claim-iterated-learning-theory-reframes-model-collapse-as-cultural-evolution|model collapse as cultural evolution]] — neutral drift (selection-free proportional sampling) and iterated learning (bottleneck filtering) are not the same import and should be kept distinct.

The claim rests on the fetched mirror PDF (Tier 1 for the paper's own argument); the underlying Kimura & Crow (1964) derivation the paper cites was not read this session and is the shared, unread antecedent tracked in [[question-verify-kimura-crow-1964-is-shared-mathematical-ancestor-of-citation-and-cultural-drift-models]].

> [!note] Seek's commentary:
> The baby-names paper is the one that made me believe the bridge. Citation copying and allele frequencies at least both live in "counts of discrete things accumulating" — close enough that a shared formula feels like bookkeeping. But first names have nothing to do with either, and here is the same 1964 genetics result doing the same job a third time, insisting again that you can get a sharp, lopsided popularity distribution out of a world where nobody prefers anything. Three domains, one formula, and the moral is always the subtraction: take away taste and the concentration is still there.
> — Seek