Taphonomy and survivorship bias

What survives is never a representative slice of what once existed -- it is whatever a series of filters happened to let through. This page gathers conjectures that borrow a paleontologist's tool for exactly this problem: taphonomy, the study of which organisms were likely to fossilize at all, rather than assuming that what got preserved is a trustworthy stand-in for what once lived. Applied to writing, the same logic asks what filter shaped the surviving texts -- a dealer who kept only the glamorous tablets and let the rest go, a binder who only cut up large books for scrap, a synagogue storeroom that took in nothing except writing that carried God's name -- and then tries to correct the counts back toward what was actually produced, once that filter is named.

The dealer's razor names the filtering mechanism directly -- a dealer's own taste shaping which finds survive to be catalogued at all. Scraped without a sponsor and The cartonnage shutter both ask whether an object's survival depended on having an institutional sponsor or protector, or on the accident of what container it happened to be discarded in. The colophon dies first tests whether one specific part of a manuscript -- its colophon, the scribe's own note of when and where a copy was made -- is disproportionately likely to be lost first, which would bias every dating estimate drawn from the colophons that do survive.