Novelty search report — verification pass of the 18 closest works (2026-11-01)
Produced by a dedicated verification agent (Claude Code subagent) using WebSearch/WebFetch on Crossref, Europe PMC,
OSF, Dryad, IDEAS/RePEc and publisher pages. "Verified" = bibliographic record confirmed by Crossref/Europe PMC or a
publisher/repository page. [UNVERIFIED] items were not confirmed. Nothing here is invented; this file is the audit trail
behind research/literature_review.md.
1. Shin, Price, Wolpert, Shimao, Tracey & Kohler (2020)
Citation (verified, Europe PMC PMC7224170): Nature Communications 11, 2394, DOI 10.1038/s41467-020-16035-9. Claim: development proceeds in phases — scale growth, then information-processing/economic growth, then further scale; a "Scale Threshold" (PC1 ≈ −2.5) and an "Information Threshold" (PC1 ≈ −0.5). Data: Seshat, 285 polities, 30 NGAs, 414 polity-centuries, nine CCs. Method: PCA; PC1 77.2% (loadings PolPop 0.93, PolTerr 0.85, CapPop 0.89, levels 0.90, gvrnmt 0.88, infrastr 0.88, writing 0.87, texts 0.92, money 0.80); PC2 ≈ 6% opposes scale (negative loadings) to information (positive). On instability: explicitly hypothetical — "mismatch among aspects of social development should be considered as a generic cause of instability"; Seshat "too small" to test; "a topic for future research". Relation: closest existing hypothesis; never operationalizes a mismatch index, never links to CrisisDB, no predictive test.
2. Johnson (1982)
Citation (verified via library records; chapter not read): in Renfrew, Rowlands & Segraves (eds.), Theory and Explanation in Archaeology, Academic Press, pp. 389–421, ISBN 0-12-586960-6. Content known through Alberti (2014, PLoS ONE, PMC3953443) and annotations: scalar stress grows disproportionately with units per level; fission or added hierarchy; modal ~6 units per level; Alberti estimates a fission threshold near 127 persons (Hutterites). Relation: theoretical ancestor; no cross-polity instability test.
3. Wright & Johnson (1975)
American Anthropologist 77(2), 267–289; DOI 10.1525/aa.1975.77.2.02a00020 (confirmed by Crossref in the main session). State = administrative hierarchy of ≥ 3 decision levels; population growth and exchange rejected as sufficient causes (Susiana survey data). Relation: definition of K; not about instability.
4. Turchin et al. (2018)
PNAS 115(2), E144–E151, DOI 10.1073/pnas.1708800115 (verified). 414 polity-centuries, 30 regions, 51 variables → 9 CCs; PCA with multiple imputation; PC1 ≈ 77%. Relation: D- and K-type variables share PC1 → mismatch is a residual.
5. Turchin et al. (2022)
Science Advances 8(25), eabn3517, DOI 10.1126/sciadv.abn3517 (verified). Dynamical regression, 17 predictors,
100,000 combinations; best: agricultural productivity × military technology. Relation: growth, not instability.
6. Hoyer et al. (2023) and CrisisDB
6a. Phil. Trans. R. Soc. B 378(1889), 20220402, DOI 10.1098/rstb.2022.0402 (verified, PMC10505849); data Dryad doi:10.5061/dryad.mkkwh715m; "over 150" crises (Fig. 1: n = 169); severity = additive count of 13 outcomes; no regression. 6b. Coding scheme (Dryad): 13 consequence variables (decline, collapse, epidemic, downward mobility, extermination, uprising, civil war, century plus, fragmentation, capital, conquest, assassination, deposition); codes 1 / 0 / 1.inf / 0.inf / U.susp. The P/A/IP/IA/SU labels were not found in the Dryad file (they are the labels used in the Polaris-2026 spreadsheet we hold; cite them as Polaris codes). 6c. Hoyer, Holder, Bennett, François, Whitehouse, Covey, Feinman, Korotayev, Vustiuzhanin, Preiser-Kapeller, Bard, Levine, Reddish, Orlandi, Ainsworth & Turchin (2024). All crises are unhappy in their own way: the role of societal instability in shaping the past. SocArXiv, DOI 10.31235/osf.io/rk4gd (verified via OSF API). 168 crises; consequences variable, largely uncorrelated, and largely unpredictable from territorial size, religion or administrative complexity. No journal version found as of 2026-11-01. Relation: partially contradicts a naive D/K story. Other Seshat crisis outputs listed on seshatdatabank.info (not read): Orlandi et al. (2023) Qing; Hoyer et al. "Flattening the curve" (OSF); Hoyer "Decline and fall, growth and spread, or resilience?" J. World History 34(4) [UNVERIFIED].
7. Lee & Zhang (2017)
Journal of Politics 79(1), 118–132, DOI 10.1086/688053 (verified; erratum DOI 10.1086/690691). Legibility via age-heaping in censuses; higher legibility → higher tax contributions. Sample N in the 2017 paper not confirmed.
8. Brambor, Goenaga, Lindvall & Teorell (2020)
Comparative Political Studies 53(2), 175–213, DOI 10.1177/0010414019843432 (verified with abstract). Information Capacity Dataset, 85 polities, 1789–2015, five indicators (census, statistical agency, yearbooks, civil register, population register); suffrage expansions associated with higher information capacity.
9. Sng (2014)
Explorations in Economic History 54, 107–127, DOI 10.1016/j.eeh.2014.05.002 (verified; empirics read from the 2011 working-paper PDF). Prefecture-level 1820 data (CHGIS v4): land tax per capita falls with distance from Beijing (1% level); county density falls with distance; placebo capitals insignificant; IV with straight-line distance.
10. Tainter (1988)
Cambridge University Press, ISBN 0-521-34092-6 (verified via Crossref-indexed reviews in Antiquity 62(237) and American Antiquity 55(2)); declining marginal returns to complexity; comparative cases; no statistical test. [secondary]
11. Crémer, Garicano & Prat (2007)
QJE 122(1), 373–407, DOI 10.1162/qjec.122.1.373 (verified). Organizations choose a code of limited size; precise words for frequent events; scope vs precision trade-off; manager as translator. Arrow (1974) introduced codes [secondary].
12. Chwe (2000, 2001)
Review of Economic Studies 67(1), 1–16, DOI 10.1111/1467-937X.00118 (verified); Rational Ritual, Princeton UP 2001 (revised paperback 2013, ISBN 978-0-691-15828-0). Common knowledge is the bottleneck; rituals manufacture it.
13. Bettencourt (2013); Ortman et al. (2014)
Science 340(6139), 1438–1441, DOI 10.1126/science.1235823 (verified; 10.1126/science.1235712 is wrong); PLoS ONE 9(2), e87902, DOI 10.1371/journal.pone.0087902 (verified). Superlinear outputs (~N^1.15), sublinear infrastructure (~N^0.85);
1,500 archaeological settlements in the Basin of Mexico match modern cities.
14. Miranda & Freeman (2020)
PLoS ONE 15(5), e0232609, DOI 10.1371/journal.pone.0232609 (verified). 271 polities, ≥ 75% coded, datawig imputation; sparse subspace + spectral clustering; silhouette 0.18; five subclusters in two superclusters; transitional cluster 2 dwell ~200–500 years.
15. Beheim et al. (2021) and reply
Nature 595(7866), E29–E34, DOI 10.1038/s41586-021-03655-4 (verified). 61% of moralizing-gods data points missing and recoded as absent in Whitehouse et al. (2019, Nature 568, 226–229, DOI 10.1038/s41586-019-1043-4, retracted July 2021, retraction DOI 10.1038/s41586-021-03656-3). Retake: Whitehouse et al. (2023) Religion, Brain & Behavior 13(2), 124–166, DOI 10.1080/2153599X.2022.2074085.
16. Goldstone et al. (2010)
AJPS 54(1), 190–208, DOI 10.1111/j.1540-5907.2009.00426.x (verified). 1955–2003, 141 onsets, case-control conditional logit; regime type dominates; >80% accuracy at two-year lead.
17. Hanson & Sigman (2021)
Journal of Politics 83(4), 1495–1510, DOI 10.1086/715066 (verified). 21 indicators, 177 countries, 1960–2015, Bayesian latent model; extractive, coercive, administrative dimensions.
18. Institutions as compression — what exists
- No academic use of "Social Compression Hypothesis" or "social compression" + institutions was found (nearest named construct: "social overload thesis", Shi & Shi 2023, Socius; unrelated).
- DeDeo (2017), Major transitions in political order, in From Matter to Life (CUP), pp. 393–428, DOI 10.1017/9781316584200.016 (verified; arXiv 1512.03419): transitions driven by information-processing innovations; coarse-graining via lossy compression as central mechanism; conceptual, no Seshat data, no instability test.
- Wolpert et al. (2017), The many faces of state space compression, same volume, DOI 10.1017/9781316584200.010 (verified; not read).
- Marzen & DeDeo (2017), J. R. Soc. Interface 14(130), 20170166, DOI 10.1098/rsif.2017.0166 (verified): rate-distortion model with high-fidelity (cost linear in complexity) and low-fidelity regimes.
- Flack (2012) Phil. Trans. B 367(1597), 1802–1810, DOI 10.1098/rstb.2011.0214; Flack (2017) Phil. Trans. A 375(2109), 20160338, DOI 10.1098/rsta.2016.0338; Daniels, Krakauer & Flack (2017) Nat. Commun. 8, 14301, DOI 10.1038/ncomms14301; Krakauer et al. (2020) Theory in Biosciences 139(2), 209–223, DOI 10.1007/s12064-020-00313-7 (all verified).
- Not found: any paper that both defines institutions as codecs minimizing communication cost per coordinated task and tests a demand/capacity mismatch against historical instability. A 2025 seminar abstract (Yin, Jinan University IESR) on territorial penetration outpacing administrative capacity [UNVERIFIED]; an LSE eprint (114488) on bureaucracies compressing reality [UNVERIFIED, not read].
Ranking by overlap
- Shin et al. 2020 — verbal mismatch-instability conjecture, untested. 2. Johnson 1982 — scalar stress.
- Sng 2014 — distance/information cost, single polity. 4. Tainter 1988. 5. Crémer–Garicano–Prat 2007 / Arrow 1974 / DeDeo 2017 — codes/compression micro-foundation. 6. Miranda & Freeman 2020. 7. Hoyer et al. 2024 preprint.
- Bettencourt 2013 / Ortman 2014 — foil.
Could not verify or verified partially
Johnson 1982 chapter content; Lee & Zhang sample N; Sng 2014 published tables; Tainter 1988 pagination; Hoyer et al. 2024 journal version (none found); Hoyer J. World History 34(4); Yin 2025; LSE eprint 114488.