Gary M. Feinman, Gabriela Cervantes Quequezana, Adam Green, Dan Lawrence, Jessica Munson, Scott Ortman, et al. (2025), Proceedings of the National Academy of Sciences (PNAS) 122(16), e2400698121. Series “The Global Dynamics of Economic Inequality over the Long Term” — DOI-10.1073/pnas.2400698121
Key takeaways
- Across 1,176 archaeological sites (47,019 houses), Gini coefficients of house size are only weakly predicted by the number of political-hierarchy levels (r² ≈ 0.22–0.27) or by polity population, and do not rise linearly or immediately with farming or urbanism. The collective-versus-autocratic character of governance explains about as much variance as hierarchy does, and the highest inequality appears at autocratic apex sites in multi-level polities.
Seeded from the abstract and a skim, before the reading in NOTE-049. What follows is what the work says about itself.
Established “grand narratives” tie inequality in large societies to roughly linear trends: domestication, then population growth or step-ups in hierarchical scale, with reversals only after catastrophes or collapse. These narratives were built mostly from historical or ethnographic snapshots and pay little attention to institutions or agency. The authors use quantitative, time-anchored archaeological and ethnographic house-size data from several world regions, most of which went through urbanization and hierarchy formation. Broad temporal patterns exist, but regional inequality trends are not linear or uniform, and they are not triggered immediately or mechanically by Malthusian dynamics or growth in scale.
Standing in the record
Filed from the survey of 2026-09-25 of work the anthology set aside as out of scope (tier B): 270 seconds of active reading over 9 sessions in the papers-feed tracker. It was filed Deferred, unread. NOTE-049 is the close reading of 2026-09-25, and it placed the work: Active — worth reading as the widest quantitative record of premodern wealth disparity and a clear rebuttal of deterministic scale-drives-inequality narratives, a negative claim the scatterplots support well. Its positive claim that governance “matters” rests on bivariate r² from an expert-coded index whose “architecture” dimension overlaps with the house-size outcome, and should be read as suggestive.
When seeded, before the reading in NOTE-049, the entry gave a priority for a deeper reading: medium — 9 sessions suggest sustained interest, but it is waiting on the social-science topic decision. The skim captures the main quantitative claims, and what remains is the SI methods.
What the seeding asked a deeper reading to check (NOTE-049 is that reading):
- A large-N test of an influential narrative (inequality as the inevitable result of scale), with the authors arguing that institutions and governance mediate the outcome.
- Check the governance coding: it was done by regional experts, apparently at the same time as the other variables. Check also the “Polity-Scale Effect” correction and how robust the r² values are to sample choice.
- The paper is part of a PNAS series. The companion papers, for example on land-limited versus labor-limited production, may be the better entry point.
Access when seeded: Crossref (online 2025-04-14) and the full text as Europe PMC XML (PMC12037030, open access), which I read: significance statement, abstract, “Approach and Methods”, “Empirically Assessing Grand Narratives”, “Governance Matters”, “Conceptual Implications” and figure captions. The PNAS PDF returned 403. The SI appendix was not read. The journal records it as received 12 Feb 2024 and accepted 12 Jun 2024.