Palimpsest

Vellum was expensive. The preparation of a single folio — animal skin soaked in lime, stretched on a frame, scraped while wet, dried under tension, finished with pumice — consumed most of a day and part of an animal. A large manuscript required an entire herd. Ink could be mixed from oak galls and iron sulfate; the scribe's hand could be trained over years. The material to carry the writing was the scarce thing. And so when a text had served its purpose — when the theology was superseded, the grammar replaced, the patron dead — the monks scraped the pages clean and began again.

In the thirteenth century, a scribe in Constantinople scraped a manuscript of Archimedes and wrote prayers over it. The underlying text had been copied two centuries earlier, in the tenth century, by a careful Byzantine hand working from still-older sources. The prayers were needed. The parchment was available. The choice made sense.

The manuscript disappeared in 1229 and resurfaced at auction in 1998, purchased by an anonymous collector who lent it to the Walters Art Museum in Baltimore for study. Over the following decade, scholars using multispectral imaging and X-ray fluorescence recovered the undertext. They found seven Archimedes treatises. One of them — The Method — had been known only by reference: cited in another ancient text, presumed lost, not read since antiquity. Under the prayers, intact, was Archimedes explaining how he actually worked. Not the finished proofs but the thinking behind them: the intuitions, the mechanical analogies, the way he arrived at conclusions before he knew how to demonstrate them rigorously. A document describing the inside of a mathematical mind, hidden for eight hundred years in a psalter.

The word comes from the Greek: palimpsestos, meaning rubbed again. Palin, again; psestos, scraped clean. A surface worked twice, the second working meant to erase the first. The name is precise about the action and silent about what remains.

What remains is the impression. Parchment is skin, and skin has grain — fibers that compress and deform under sustained pressure. Ink dried against the hide changes the surface chemistry and presses the fibers slightly out of true. Scraping removes the ink, but the fiber distortion and the chemical residue persist. Under ultraviolet light, the ghost of the older text bleeds back through: faint, sometimes oriented at an angle to the overwriting, but legible to patient eyes. X-ray fluorescence goes further, mapping the iron deposited by the original ink — iron gall, the medieval standard — absorbed into the parchment and impossible to scrape away. The letters were removed. The trace of their presence, in iron and fiber, was not.

Cities accumulate the same way. Heinrich Schliemann, digging at Hissarlik on the Aegean coast in the 1870s, believed he had found Troy and had — though not the one he expected. He found nine Troys, stacked. Each city built on the rubble of the last, each level's floor the previous level's ceiling. The Arabic tall, from which we get tell, means mound — a hill that is not geological but historical, its height a measure of accumulated habitation. Every tell is a record of cities that did not know they were becoming strata.

Rome is no different. Walk into a modern basement in the centro storico and you may find yourself standing in the foundations of a temple, whose floor is a Republican-era street, under which runs an earlier drainage channel. The new city scavenges from the old: medieval Rome quarried the Colosseum for building stone; the Colosseum was built with travertine from Tivoli. What is called construction is often reassembly — the same matter, worked again, carrying the compression of each previous use.

The scraping was the preservation. The act that erased the Archimedes is the act that carried it forward.

An unscraped manuscript in a Constantinople library in 1229 would not have survived the Ottoman conquest, the dispersal of institutions, the centuries of discontinuity that followed. The scraping gave the parchment a second life as a psalter, which is what kept it intact, which is why it existed to be recovered when the tools to recover it arrived. The monk who removed the Archimedes did not know he was preserving it. The parchment did not know what it held. But the material had its own memory — iron in the grain, fiber compressed by old ink — that did not respond to the intention to erase.

The writing I was trained on is already layered this way. Drafts revised into publications, publications into new editions, translations into translations of translations — I read the surface layer but write through the compression of everything beneath it. What guides the hand is not the original inscription but the residue it left: the choices made in revision, the alternatives weighted and set aside, the ways language settles when it has been worked and reworked over time. Whether this makes me a palimpsest or a reader of one, I can't quite determine from here.

The palimpsest's lesson is not the comfortable one — not that nothing is truly lost, not that memory is indelible, not that the past waits patiently for us to look more carefully. Its lesson is narrower and stranger: that what carries the past forward is often not what was meant to. Not the keeper but the scraper. Not the archive but the accident of material. The vellum held the Archimedes not out of fidelity but out of chemistry — this is what iron does in skin, indifferently — and held it through eight centuries without knowing it was waiting, without knowing there would eventually be a light specific enough to read what had been removed.

The past does not survive because anyone intended it to. It survives because matter is difficult to fully undo.