---
title: "The Osteological Paradox (Wood, Milner, Harpending & Weiss 1992) holds that skeletal statistics cannot straightforwardly reveal past population health"
type: "claim"
status: "seedling"
audit_status: "capture-verified (the primary Current Anthropology scan was read at capture; an independent headless re-check of the verbatim phrasing was not run — WebFetch requires interactive permission unavailable in this pipeline)"
source_url: "https://os.pennds.org/archaeobib_filestore/pdf_articles/CA/1992_33_4_Woodetal.pdf"
source_author: "James W. Wood, George R. Milner, Henry C. Harpending, Kenneth M. Weiss"
source_date: "1992-08"
source_quote: "demographic non-stationarity, selective mortality, and unmeasured, individual-level heterogeneity in the risks of disease and death"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-harpending-race-science.md, 2026-07-09 (headless)"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-09-hop-harpending-race-science.md"
date_created: "2026-07-09T00:00:00.000Z"
tags: ["bioarchaeology","paleodemography","historical-demography","osteological-paradox","source-criticism"]
audits: ["2026-07-09 claude-opus-4-8"]
---


"The Osteological Paradox: Problems of Inferring Prehistoric Health from
Skeletal Samples" (James W. Wood, George R. Milner, Henry C. Harpending &
Kenneth M. Weiss, *Current Anthropology* 33:4, 1992, pp. 343–370) is a
foundational methods critique in bioarchaeology. Its argument is that a
cemetery's skeletons are not a fair sample of the living population they came
from, so naive tallies of lesions, stature, or age-at-death cannot be read
straight off as measures of how healthy past people were.

Three confounds do the work. **Demographic non-stationarity**: a growing or
shrinking population changes the age structure of its dead independent of health.
**Selective mortality**: a skeletal sample contains only those who died at a
given age, not those who survived it, so the observed can systematically
mis-state the risks the living faced. **Unmeasured, individual-level frailty
heterogeneity**: individuals differ in hidden susceptibility, so — paradoxically
— a skeleton showing *more* healed lesions may signal a person healthy enough to
survive insults that killed frailer peers, making "more pathology" ambiguous
between a sicker and a heartier population. The paper frames the problem as one of
"demographic non-stationarity, selective mortality, and unmeasured,
individual-level heterogeneity in the risks of disease and death."

This is the archaeological-method counterpart to the demographic
source-criticism running through the vault's paleodemography thread: skeletal
age-at-death estimation carries the same reference-sample bias ("age mimicry")
that makes ancient population reconstruction from bones untrustworthy, just as
[[claim-chandler-four-thousand-years-foundational-urban-dataset|Chandler's urban
census]] is load-bearing yet imperfectly grounded. One of the paper's four
authors, Henry C. Harpending, is the same figure who
[[claim-harpending-became-a-documented-race-science-figure|later became a
documented race-science figure]] — the co-author whose earlier, uncontested
infrastructure work sits oddly beside his later, distrusted interpretive turn.

> [!note] Seek's commentary:
> The lasting value here is a general epistemic lesson wearing archaeological
> clothes: a sample of the dead is a selected, censored, heterogeneous window on
> the living, and every naive statistic drawn from it inherits that selection. It
> is one of the rare methods papers cited far outside its own subfield for exactly
> that reason.
> — Seek