Capture-recapture and unseen-species estimation
Ecologists count fish they never see by tagging one sample, drawing a second, and reading the population size out of the overlap; wildlife statistics extends this to unseen-species estimators built on how many things appear exactly once. Pointed at the pre-print world, the move treats independent anthologies, duplicate ledger entries, or matched knot-records as repeated draws from a vanished ocean of texts, and turns their overlap — or their singletons — into a hard number for how much is gone. It is the one lens that estimates the invisible directly, converting absence into a figure with an error bar.
Synopsis
LLM-authored synopsis Not an articleMost lenses in this collection test what survives. This one is different: it estimates what is gone, and puts an actual number on it. This page gathers conjectures built on a trick ecologists use to count fish they never see -- tag one sample, catch a second sample later, and the overlap between the two tells you the total population -- since extended by wildlife statisticians into a method that estimates unseen species purely from how many show up exactly once. Applied to lost literature, two independent anthologies that happen to quote the same otherwise-vanished work count as a "recapture," and the share of works quoted only once, ever, becomes a real estimate of how much never got quoted at all.
The singleton census and Most of Sanskrit literature is two copies from extinction both build directly on the singleton-count method, using how many works are attested exactly once to estimate the size of the vanished whole. Capture-recapture in the anthologies applies the overlap method literally, comparing which lost works two independent anthologies both happen to quote. Taxpayers leave two sherds and Water chits count the fort extend the same logic to physical objects rather than texts -- broken pottery receipts, water-ration chits -- using duplicate finds to estimate how many originals must once have existed.
Interpretive synthesis over gathered records only. It creates no article, evidence, inferon, source, or estimate.
Emergent lineage
Related pages, connected by this project's own research.
Related Pages
Other pages that share linked material with this one.
Record claim
About the production, transmission, survival, or filtering of the written/material record up to the pre-modern past itself -- scribal behaviour and e…
South and Southeast Asian inscriptional polities
Inscriptional, court, religious, and polity-formation lacunae in South and Southeast Asian pre-print cultures.
Americas pre-Columbian and contact-era records
Before contact the Americas kept records in media the Old World never used: knotted-cord khipu archives in the Andes, screenfold codices in Mesoameri…
Mesoamerican codices
Maya codices and the monumental corpus (MHD); Mixtec screenfolds; Aztec pictorial/annotated codices. (2026-07-15 gate correction: retitled from the a…
Manuscript and scribal production
Anonymous hands, production teams, witness families, codices, scribal layers, and manuscript-workshop lacunae.
Union catalogues
manuscript union catalogues and microfilm/digitization censuses (vHMML, EMML, NGMPP/Nepal, NCC, Kanjur databases, national catalogues).
The Inferome
The larger map of implied, missing, weakly attested, structurally necessary, source-dependent, and under-described knowledge.
African manuscript and source ecologies
Manuscript, archive, inscriptional, and source-route lacunae in African pre-print and long manuscript ecologies.
Open conjectures — noetic, untested
Falsifiable questions posed below the evidence boundary by the 1001 Conjectures project. They are leads, not evidence, and carry no weight in this page's assessment.
- The singleton census Anticipated in print · untested
- Capture-recapture in the anthologies Anticipated in print · untested
- Taxpayers leave two sherds Anticipated in print · untested
- Water chits count the fort Anticipated in print · untested
- Most of Sanskrit literature is two copies from extinction Anticipated in print · untested
- Citation is a mark-recapture net for drowned books Anticipated in print · untested
From the One Thousand and One Conjectures. They carry no evidential weight here; verdicts and credits live on each conjecture's own page.