arxiv: 2604.10957 · v1 · submitted 2026-04-13 · 🧬 q-bio.PE · cs.AI· cs.CL· cs.CY

Recognition: unknown

A molecular clock for writing systems reveals the quantitative impact of imperial power on cultural evolution

Hiroki Fukui

Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:28 UTC · model grok-4.3

classification 🧬 q-bio.PE cs.AIcs.CLcs.CY

keywords writing systemsmolecular clockcultural evolutionphylogeneticsscript replacementimperial powercultural replicatorsdivergence times

0 comments

The pith

Writing systems evolve at a steady molecular clock rate disrupted by imperial interventions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Researchers assembled a database covering 300 writing and notation systems across 5,400 years, encoded with 50 binary structural characters, to reconstruct their historical relationships. They found that scripts change according to a consistent substitution rate under a strict clock model, producing reliable divergence estimates. Strong political interventions cause scripts to deviate from these expected times in proportion to intervention intensity. This pattern indicates that external power can selectively accelerate or redirect structural evolution rather than simply speeding all change uniformly. The analysis also identifies a limit on new script creation where writing is already present and higher extinction odds under colonial contact.

Core claim

Scripts exhibit a detectable molecular clock. The best-fitting model yields a substitution rate of 0.226 substitutions per character per millennium. Political interventions break this clock, as deviation from expected divergence times correlates with intervention intensity, and per-character rate analysis shows that interventions selectively rewrite deep structural features rather than merely accelerating overall change. Thirty major script replacement events can be ranked by destructive impact.

What carries the argument

The Mk+Gamma strict molecular clock model applied to 50 binary structural characters across 300 scripts, which estimates divergence times and identifies deviations linked to political interventions.

If this is right

Phylogenetic methods can date script divergences using the estimated clock rate.
Intervention strength can be quantified through measurable departures from expected script divergence.
Certain empires produced more script extinctions among contacted systems than others.
Independent invention of new scripts is suppressed once writing exists in a region.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same clock-based approach could measure external influences on other cultural traits that leave structural records.
Modern globalization might act as a large-scale intervention that homogenizes or redirects cultural replicators in predictable ways.
Cultural evolution appears coupled to political history through measurable effects on trait divergence and survival.

Load-bearing premise

The 50 binary structural characters are homologous, independent, and sufficient to reconstruct accurate relationships across all scripts, and intervention intensity can be scored objectively without circularity to the character data or dates.

What would settle it

A set of scripts with no recorded political interventions that still display large, systematic deviations from the divergence times predicted by the 0.226 substitution rate.

Figures

Figures reproduced from arXiv: 2604.10957 by Hiroki Fukui.

**Figure 2.** Figure 2: Figure 2. Selected phylogenetic lineages of writing systems. Line color indicates transmission [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗

**Figure 3.** Figure 3: Figure 3. Molecular clock of writing systems. (a) All 246 parent-child pairs ( [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗

**Figure 4.** Figure 4: Figure 4. Per-character substitution rates for 50 structural features, ranked from most con [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗

**Figure 5.** Figure 5: Figure 5. Top 20 script replacement events ranked by integrated destruction score. Color [PITH_FULL_IMAGE:figures/full_fig_p018_5.png] view at source ↗

**Figure 6.** Figure 6: Figure 6. Number of living writing systems through time ( [PITH_FULL_IMAGE:figures/full_fig_p019_6.png] view at source ↗

**Figure 7.** Figure 7: Figure S1. Sensitivity analysis of destruction score rankings. Box plots show rank distribution [PITH_FULL_IMAGE:figures/full_fig_p027_7.png] view at source ↗

**Figure 8.** Figure 8: Figure S2. MCMC trace plots for Bayesian clock analysis. Top: strict clock (log q and log [PITH_FULL_IMAGE:figures/full_fig_p027_8.png] view at source ↗

**Figure 9.** Figure 9: Figure S3. Posterior distribution of the Gamma shape parameter [PITH_FULL_IMAGE:figures/full_fig_p028_9.png] view at source ↗

**Figure 10.** Figure 10: Figure S4. BIC comparison across four models: Mk strict, Mk relaxed, Mk+ [PITH_FULL_IMAGE:figures/full_fig_p028_10.png] view at source ↗

read the original abstract

Writing systems are cultural replicators whose evolution has never been studied quantitatively at global scale. We compile the Global Script Database (GSD): 300 writing and notation systems, 50 binary structural characters, and 259 phylogenetic edges spanning 5,400 years. Applying four methods -- phenetics, cladistics, Bayesian inference, and neural network clustering -- we find that scripts exhibit a detectable molecular clock. The best-fitting model (Mk+Gamma strict clock) yields a substitution rate of q = 0.226 substitutions/character/millennium (95% CI: 0.034-1.22; Delta BIC = -4.1 versus relaxed clock; Delta BIC = -1,364.7 versus Mk without rate variation). Political interventions break this clock: deviation from expected divergence times correlates with intervention intensity (Spearman rho = 0.556, p < 10^{-4}), and per-character rate analysis reveals that intervention selectively rewrites deep structural features rather than merely accelerating change (rate profile correlation rho = 0.320). We identify 30 major script replacement events and rank their destructive impact. A ceiling effect suppresses independent invention wherever writing already exists (Fisher's exact OR = 0.054, p < 10^{-6}), and colonial contact predicts script extinction (Cox HR = 5.25, p = 0.0006). The Spanish Empire extinguished the most scripts (6 of 12 contacted, 50%), followed by the Empire of Japan (3 of 9, 33.3%). Feature coding was validated by inter-rater reliability testing with two independent human coders (Cohen's kappa = 0.877; human-LLM kappa = 0.929; Fleiss' kappa = 0.911).

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper compiles a global database of 300 scripts and fits a molecular clock model that correlates with imperial interventions, but the 50 binary characters are unlikely to be independent or homologous enough for reliable phylogeny.

read the letter

The core takeaway is that this work assembles the first large-scale dataset on writing systems and reports a detectable clock rate of 0.226 substitutions per character per millennium, with political events linked to faster change and extinctions. The database itself and the model comparisons are the real additions here. They test phenetics, cladistics, Bayesian inference, and neural clustering, then tie deviations to intervention intensity with a Spearman correlation of 0.556. They also rank empires by script loss and note a ceiling effect on new inventions. That combination of scale and historical linkage is new and worth having on record. The inter-rater kappas above 0.87 show the coding was done consistently. The BIC drops are large, and the extinction hazard ratio for colonial contact comes out at 5.25. Those numbers are concrete and can be checked against the data they release. The soft spots sit in the characters and the dating. Fifty binary traits on 300 taxa is a thin matrix, and script features often spread by borrowing or converge independently, so homology and independence are questionable. The stress-test concern about inflated clock support holds up on the abstract and the described methods; nothing in the summary shows they ran character correlation tests or used dated fossils independent of the historical events they later correlate against. The phylogenetic edges are mentioned but not explained in enough detail to rule out circularity. This is for researchers who already work on cultural phylogenetics or quantitative history of technology. A reader who wants raw data on scripts or a first pass at clock-like models in culture will find usable material. It is coherent enough and the questions are large enough that it deserves a serious referee, mainly to pressure-test the character matrix and the independence of the dating. I would send it out rather than desk-reject, with the expectation that the homology issue will need direct response.

Referee Report

3 major / 3 minor

Summary. The paper compiles the Global Script Database (GSD) of 300 writing and notation systems scored on 50 binary structural characters, yielding 259 phylogenetic edges over 5,400 years. Applying phenetic, cladistic, Bayesian, and neural-network methods, it reports that scripts follow a detectable molecular clock, with the Mk+Gamma strict clock as the best model (q = 0.226 substitutions/character/millennium, 95% CI 0.034–1.22; ΔBIC = −4.1 vs. relaxed clock, ΔBIC = −1,364.7 vs. equal-rates Mk). Political interventions are shown to disrupt the clock via correlation between deviation from expected divergence times and intervention intensity (Spearman ρ = 0.556, p < 10^{-4}), with additional results on selective rate changes, 30 replacement events, a ceiling effect on independent invention (OR = 0.054), and colonial contact as a predictor of extinction (Cox HR = 5.25). Inter-rater reliability for character coding is reported as high (Cohen’s κ = 0.877, human-LLM κ = 0.929, Fleiss’ κ = 0.911).

Significance. If the central claims hold, the work supplies the first global-scale quantitative demonstration of clock-like evolution in a cultural replicator and links imperial power to measurable acceleration or disruption of that process. The scale of the GSD, the multi-method convergence, and the explicit validation of character coding are genuine strengths that could anchor future comparative studies of cultural phylogenetics. The reported rate, intervention correlations, and extinction models are falsifiable and therefore potentially high-impact for historical linguistics and cultural evolution research.

major comments (3)

[Methods] Methods (character matrix construction): The 50 binary structural characters are load-bearing for all phylogenetic and clock inferences, yet the manuscript provides no explicit test or argument for their homology and independence. Script traits frequently arise by diffusion (e.g., alphabetic features from Phoenician/Greek sources) or convergence (abjad vs. alphabet distinctions), violating the Mk model’s assumptions. High inter-rater kappas confirm reproducibility but not phylogenetic validity; with only 50 characters for 300 taxa the matrix is sparse, and undetected correlations can artifactually inflate ΔBIC values (especially the −1,364.7 gap) and produce spurious clock-like behavior.
[Results (clock model)] Phylogenetic dating and clock fitting: The 259 edges are stated to span 5,400 years, but the text does not detail how these dates were obtained independently of the clock model itself or what controls were applied for sampling bias and incomplete coverage. Without such independence, the strict-clock preference (ΔBIC = −4.1) and the subsequent intervention-deviation correlation cannot be cleanly separated from the dating procedure used to calibrate the rate q.
[Results (intervention effects)] Intervention analysis: The claim that political interventions selectively rewrite deep structural features (rate-profile ρ = 0.320) and that intervention intensity predicts clock deviation rests on the same character matrix and dated tree. The manuscript does not report sensitivity checks that remove putatively non-neutral characters or that use alternative historical scorings of intervention intensity; therefore the causal interpretation remains vulnerable to the same homology and independence concerns raised above.

minor comments (3)

[Abstract / Methods] The abstract and methods should state the fraction of the 300 systems that have complete character data and the number of edges whose dates rely on external historical sources versus inferred from the clock.
[Results] Notation for the 95% CI on q should be clarified (is it a credible interval from the Bayesian run or a profile-likelihood interval?) and the exact BIC calculation (including the number of free parameters) should be given in a table.
[Figures] Figure legends for the rate-profile and extinction analyses should explicitly list the number of characters or events entering each correlation.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which identify key areas for strengthening the methodological foundations of our phylogenetic analysis of writing systems. We respond point-by-point to the major comments below, offering clarifications based on the existing analyses and committing to targeted revisions that enhance transparency and robustness without altering the core findings.

read point-by-point responses

Referee: [Methods] Methods (character matrix construction): The 50 binary structural characters are load-bearing for all phylogenetic and clock inferences, yet the manuscript provides no explicit test or argument for their homology and independence. Script traits frequently arise by diffusion (e.g., alphabetic features from Phoenician/Greek sources) or convergence (abjad vs. alphabet distinctions), violating the Mk model’s assumptions. High inter-rater kappas confirm reproducibility but not phylogenetic validity; with only 50 characters for 300 taxa the matrix is sparse, and undetected correlations can artifactually inflate ΔBIC values (especially the −1,364.7 gap) and produce spurious clock-like behavior.

Authors: We agree that explicit arguments for homology and independence are essential for validating the Mk model. The 50 characters were chosen as deep structural traits (e.g., directionality, syllabicity, and logographic elements) drawn from established classifications in historical linguistics to minimize diffusion effects, but the manuscript does not include formal tests. In revision, we will add a dedicated Methods subsection providing homology arguments with citations to linguistic sources, pairwise character correlation analyses, and robustness checks by refitting models after excluding potentially convergent traits (e.g., alphabetic features). These additions will confirm that the large ΔBIC gap and clock-like signal are not artifacts of undetected correlations. The convergence across four independent methods (phenetics, cladistics, Bayesian, neural networks) and high inter-rater kappas provide additional support for the phylogenetic validity. revision: yes
Referee: [Results (clock model)] Phylogenetic dating and clock fitting: The 259 edges are stated to span 5,400 years, but the text does not detail how these dates were obtained independently of the clock model itself or what controls were applied for sampling bias and incomplete coverage. Without such independence, the strict-clock preference (ΔBIC = −4.1) and the subsequent intervention-deviation correlation cannot be cleanly separated from the dating procedure used to calibrate the rate q.

Authors: The 259 edge dates were compiled independently from historical records, epigraphic data, and linguistic scholarship (detailed in the supplementary materials), separate from the clock model fitting. We will revise the main-text Methods to explicitly enumerate the sources and calibration criteria for each dated node, along with controls for sampling bias including geographic/temporal subsampling and rarefaction analyses to assess coverage effects. These independent calibrations underpin both the strict-clock model selection and the intervention-deviation correlation; we will further demonstrate separation by reporting sensitivity results using alternative date calibrations. revision: yes
Referee: [Results (intervention effects)] Intervention analysis: The claim that political interventions selectively rewrite deep structural features (rate-profile ρ = 0.320) and that intervention intensity predicts clock deviation rests on the same character matrix and dated tree. The manuscript does not report sensitivity checks that remove putatively non-neutral characters or that use alternative historical scorings of intervention intensity; therefore the causal interpretation remains vulnerable to the same homology and independence concerns raised above.

Authors: We acknowledge that additional sensitivity checks would strengthen the causal interpretation of intervention effects. In the revised manuscript, we will report new analyses that exclude characters potentially subject to convergence or diffusion and that employ alternative historical databases for scoring intervention intensity. These checks will test the stability of the rate-profile correlation (ρ = 0.320) and the clock-deviation correlation (ρ = 0.556), confirming that interventions selectively target deep features rather than merely accelerating neutral change. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected in the derivation chain.

full rationale

The paper constructs a phylogeny from 50 binary structural characters across 300 scripts using standard methods (phenetics, cladistics, Bayesian inference), then fits an Mk+Gamma strict clock model to estimate the substitution rate q and performs BIC-based model comparisons. This is conventional phylogenetic model fitting and selection, not a reduction of outputs to inputs by construction. Clock deviations are then correlated against independently sourced historical intervention data (Spearman rho), and extinction/ceiling effects are tested against external records of colonial contact and empires. Inter-rater reliability (Cohen's kappa, Fleiss' kappa) validates character coding separately from the clock analysis. No self-citations, ansatzes, or uniqueness theorems are invoked as load-bearing premises, and no fitted parameters are relabeled as independent predictions. The central claims remain externally falsifiable against historical benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on treating script structural traits as evolving in a clock-like manner under phylogenetic models borrowed from biology, with the database and intervention scores treated as independent inputs.

free parameters (1)

substitution rate q = 0.226 substitutions/character/millennium
Fitted parameter from the Mk+Gamma strict clock model to the 259-edge dataset.

axioms (2)

domain assumption Binary structural characters of scripts can be treated as independent, heritable traits that evolve via substitutions analogous to molecular sequences.
Invoked to justify application of phenetics, cladistics, Bayesian clock models, and neural clustering to cultural data.
domain assumption The 259 phylogenetic edges and associated divergence times accurately reflect historical relationships independent of the fitted clock rate.
Required for testing deviations caused by political interventions and for ranking replacement events.

pith-pipeline@v0.9.0 · 5633 in / 1803 out tokens · 98892 ms · 2026-05-10T16:28:07.524993+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

18 extracted references

[1]

Bentz, C., & Dutkiewicz, E. (2026). Humans 40,000 y ago developed a system of conventional signs.Proceedings of the National Academy of Sciences, 123(9), e2520385123

2026
[2]

(2003).Writing Systems: An Introduction to Their Linguistic Analysis

Coulmas, F. (2003).Writing Systems: An Introduction to Their Linguistic Analysis. Cambridge University Press

2003
[3]

T., & Bright, W

Daniels, P. T., & Bright, W. (Eds.). (1996).The World’s Writing Systems. Oxford University Press

1996
[4]

Felsenstein, J. (1981). Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution, 17(6), 368–376

1981
[5]

D., & Atkinson, Q

Gray, R. D., & Atkinson, Q. D. (2003). Language-tree divergence times support the Anatolian theory of Indo-European origin.Nature, 426(6965), 435–439

2003
[6]

J., Wu, C.-H., Hua, X., Dunn, M., Levinson, S

Greenhill, S. J., Wu, C.-H., Hua, X., Dunn, M., Levinson, S. C., & Gray, R. D. (2017). Evolutionary dynamics of language systems.Proceedings of the National Academy of Sciences, 114(42), E8822– E8829

2017
[7]

Hosszú, G. (2024). Validation of graph sequence clusters through multivariate analysis: application to Rovash scripts.npj Heritage Science, 12, 110

2024
[8]

E., & Raftery, A

Kass, R. E., & Raftery, A. E. (1995). Bayes factors.Journal of the American Statistical Association, 90(430), 773–795

1995
[9]

Kimura, M. (1968). Evolutionary rate at the molecular level.Nature, 217(5129), 624–626

1968
[10]

Lewis, P. O. (2001). A likelihood approach to estimating phylogeny from discrete morphological character data.Systematic Biology, 50(6), 913–925

2001
[11]

Lieberman, E., et al. (2007). Quantifying the evolutionary dynamics of language.Nature, 449(7163), 713–716

2007
[12]

Mace, R., & Holden, C. J. (2005). A phylogenetic approach to cultural evolution.Trends in Ecology & Evolution, 20(3), 116–121

2005
[13]

D., & Meade, A

Pagel, M., Atkinson, Q. D., & Meade, A. (2007). Frequency of word-use predicts rates of lexical evolution throughout Indo-European history.Nature, 449(7163), 717–720

2007
[14]

(2007).The Story of Writing

Robinson, A. (2007).The Story of Writing. Thames & Hudson

2007
[15]

Rosa, M. C. (2010). Ryukyuan writing systems and notation. [Conference paper]

2010
[16]

Rosa, M. C. (2016). Kaid ¯a glyphs in the digital humanities. InProceedings of JADH 2016

2016
[17]

Yang, Z. (1994). Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods.Journal of Molecular Evolution, 39(3), 306–314. 25 PREPRINT v5

1994
[18]

Zuckerkandl, E., & Pauling, L. (1965). Evolutionary divergence and convergence in proteins. In Evolving Genes and Proteins(pp. 97–166). Academic Press. 26 PREPRINT v5 8 Supplementary Materials Figure 7: Figure S1. Sensitivity analysis of destruction score rankings. Box plots show rank distribution across 1,000 random weight configurations (Dirichlet-sampl...

1965