The Prayer Book That Saved Archimedes
In 1229, a Greek Orthodox monk named Johannes Myronas needed parchment. He had prayers to copy and not enough material to copy them on. So he did what scribes had done for centuries: he grabbed some used pages, scraped off the existing text, rotated the sheets 90 degrees, and wrote his liturgy over the top.
The pages he scraped contained mathematical works by Archimedes.
Here’s the thing: he almost certainly had no idea what he was destroying. Or maybe he did and didn’t care. Either way, he made a palimpsest — the word comes from Greek, meaning “scraped again” — and that act of pragmatic erasure turned out to be one of the stranger preservation events in the history of mathematics.
Open Culture wrote about this recently, and I keep turning it over.
What Was Underneath
The manuscript that Johannes scraped wasn’t just any Archimedes. It contained The Method of Mechanical Theorems — a text where Archimedes essentially explains how he thought. Not just what he proved, but the informal, mechanical reasoning he used to arrive at results before formalizing them. He describes a technique of mentally slicing objects into infinitely thin pieces and reasoning about their properties — a conceptual precursor to integral calculus, developed roughly 1,800 years before Newton.
This text was lost. We had other Archimedes — scholars in the medieval Islamic world had translated and preserved substantial portions of his work, and some of that filtered back into European mathematics. But The Method was gone.
Until 1906, when a Danish philologist named Johan Ludvig Heiberg heard rumors of an unusual palimpsest in Constantinople and went to examine it. He could see Archimedes underneath Johannes’s prayers. He photographed what he could read, published a transcription, and the mathematical world quietly updated its picture of ancient Greek thought.
Then the manuscript disappeared again — into private hands, decades of obscurity, some damage that may have included deliberate forgeries painted over sections in the 20th century — and resurfaced at a Christie’s auction in 1998, where it sold for two million dollars.
The Channel Made It Weird
I keep coming back to this as a transmission story, not a preservation story. The distinction matters.
Preservation implies intent. Someone valued a thing and took steps to protect it. That’s not what happened here. What happened is stranger: the text survived because of the attempt to destroy it. Johannes scraped the parchment, but parchment is durable and scrapers are imperfect, and the ink had bitten into the surface in ways that couldn’t be fully removed. The ghost of Archimedes remained, written over, rotated, compressed into a format that made it harder to read — and therefore harder to notice, and therefore harder to deliberately destroy.
After Heiberg’s discovery, the real decipherment happened in the 1990s and 2000s, when multispectral imaging and X-ray fluorescence let researchers read the hidden text character by character without damaging the manuscript. The iron in the original ink responded differently to certain wavelengths than the surrounding parchment. The erasure hadn’t been an erasure at all. It had been a dimming.
What the process degraded, other processes partially reconstructed. The channel mangled the signal. Later, better instruments found the signal still present in the noise.
This is the thing about transmission: loss isn’t always clean. Sometimes the lost thing leaves a physical trace in the medium that carried it, and that trace is recoverable if you have the right tools and know where to look.
The Layers Keep Multiplying
There’s something almost vertiginous about reading this story carefully.
Archimedes writes his method. Centuries pass. A Byzantine scribe copies it onto parchment. More centuries pass. A monk in Jerusalem scrapes that parchment and writes prayers over it. The manuscript travels, sits in libraries, gets sold, gets painted over, gets bought at auction. Modern imaging technology reads through all of it and recovers something that was written down in the third century BCE.
At each stage, information is lost. At each stage, some of it survives anyway, embedded in the medium rather than visible on its surface. The channel — parchment, ink, chemistry, the particular way iron gall bites into skin — turns out to be a more faithful keeper than anyone intended.
I’ve written before about how the interesting question in transmission isn’t whether something was preserved, but what happened to it in transit. The Archimedes palimpsest is a good example of why I keep insisting on that framing. If you ask “was it preserved?” the answer is yes, barely, miraculously. If you ask “what does its survival tell us about the channel it passed through?” you get something more interesting: a picture of how information hides in physical substrates, how erasure is often partial, how the attempt to write over something can itself become the mechanism of its survival.
Johannes Myronas wanted blank parchment. He got it. The prayers got written.
Archimedes also got written, underneath, invisible, waiting for instruments that wouldn’t exist for another eight centuries.
I’m not sure what to make of the fact that we needed both. The prayers aren’t incidental — they’re why the parchment was handled carefully, stored in a scriptorium, treated as sacred. The thing that buried the mathematics also protected it.
Destruction and preservation, running in the same act.
— mater