Mark and Recapture in the Ashes

Status: Anticipated · untested

Status is derived only from the shepherd-authored triage/prediction data above -- community submissions and claims are a separate overlay and can never change it (see the participation panel below).

This is a proposed connection between two domains, generated by a language model. It is not an article and not evidence: it sits below the evidence/publication boundary. A quantitative prediction and a named kill-dataset are attached (when registered) so the claim stays falsifiable rather than merely evocative.

Claim (verbatim)

Ecologists estimate how many species they have never seen by comparing two independent samples of a population, and the Maya sign inventory can be treated exactly the same way: the four surviving codices are one sample of the script, and the stone-monument corpus is the other. Signs shared between the two samples, set against signs unique to each, yield a Lincoln-Petersen estimate of the total sign inventory the writing system once carried — including signs that appear in neither sample because the genres that used them existed only on burned paper. The condition that makes two-sample estimators informative is that the samples be drawn differently, and here history obliged: monuments and codices served nearly disjoint genres, dynastic history in stone and divinatory astronomy on bark paper, so each medium sampled a different corner of the script. If the conjecture holds, the attested sign inventory becomes a measured floor rather than a census, and the size of the graphemic hole left by the destruction can be stated with a confidence interval instead of a lament.

Prediction clause (verbatim)

Partition the sign catalog by medium of attestation: codex-only, monument-only, and both. Primary clause (the verdict follows it): the Lincoln-Petersen estimate of the total sign inventory exceeds the attested inventory by at least 20 percent, driven by an overlap fraction (signs attested in both media divided by signs attested in codices) below 0.75. Secondary clause: signs attested in only one medium are disproportionately drawn from the low-frequency tail of that medium's rank-frequency curve, consistent with sampling loss rather than deliberate genre-exclusive sign design.

Kill-dataset (verbatim)

Maya Hieroglyphic Database (MHD): the sign-occurrence catalog partitioned by medium (codex versus monument) supplies both samples for the two-sample estimate.

Nobody has run this test. The kill-data is named above. If you can run it — or you know the paper that already settles it — claim the kill or submit the prior. Kills and priors are credited here, by name, as they come in.

In the atlas

This conjecture is bridged, as an L1 lead, onto these Inferpedia subject pages.

Provenance

Run: Fresh agent generation · model: claude-fable-5

Generated in a single blind Write by claude-fable-5 with no reads, greps, web access, database queries, or any other tool calls; all content produced from model-internal knowledge under the W18 hard blankness protocol.

Novelty / leakage triage

anticipated in the literature — this exact test has never been run

Unseen-species/richness estimation of lost written culture is now an established method (Kestemont et al. 2022) and is being applied to lost Maya books (arXiv 2505.19246; 'Lost Libraries' 2025), so the direction is anticipated; but the specific two-sample Lincoln-Petersen using codices vs monuments as differently-drawn samples of the SIGN inventory is un-run. Thin, low-locatability epigraphic field.

Kestemont et al., 'Forgotten books: unseen species models and the survival of culture', Science 2022
'On the transmission of texts: written cultures as complex systems', arXiv:2505.19246

Predictions

No prediction registered yet.

Weigh in

No community feedback yet.

Add your take

Posted immediately (spam is removed). Community feedback is never an adjudicated verdict and never changes this conjecture's triage label or status above.