AI Solves 370-Year-Old Royalist Cipher Hidden in 17th-Century Text
Claude Fable 5.1 cracks two historical cryptograms using contextual analysis
1The Unsolved Cipher Challenge
The Cyphral Distich, a 64-number cryptogram appended to Sir Thomas Urquhart's 1653 Logopandecteision, had baffled cryptographers for centuries. Listed among Klaus Schmeh's Top 50 unsolved encrypted messages, it resisted traditional methods like frequency analysis and substitution ciphers. The puzzle consisted of two lines of 32 numbers each, with historical attempts dating back to an 1899 publication in Notes and Queries. This 370-year-old mystery represented a perfect test case for AI's contextual reasoning capabilities.
2AI's Breakthrough Approach
Claude Fable 5.1 solved the cipher in 44 minutes using 176k tokens, employing two key realizations. First, it noted the cipher's proximity to Urquhart's 32 Proquiritations - a numerically significant section the author emphasized. Second, it interpreted poetic clues about containing the author's "minde" and readers' "wishes" literally. Unlike human cryptographers who sought external cipher keys, the AI recognized the solution lay within the text itself - a paradigm shift in cryptographic methodology that leveraged the book's internal structure as the decryption key.
3The Decryption Mechanism
The solution involved mapping each number to Urquhart's Proquiritations: the i-th number corresponded to the i-th Proquiritation, with the numerical value indicating a word position. Taking the first letter of each selected word revealed a royalist prayer: "O GOD UPHOLD KING CHARLS THE SECOND AND MAKE HIM THE SUPREME RULER OF THIS LAND". The 32-letter result formed a rhyming couplet (land/and), matching the promised "distich" format and aligning perfectly with Urquhart's well-documented Royalist political leanings during the English Civil War period.
4Verification and Historical Context
The solution self-verified through multiple dimensions: the 32-letter output matched the cipher's 32-number input lines, the rhyming structure fulfilled the "distich" promise, and the pro-monarchy content aligned with Urquhart's political writings. As a committed Royalist during Cromwell's rule, Urquhart had motive to hide such messages. The AI's solution also explained why previous attempts failed - they sought external cipher alphabets rather than the internal textual reference that Urquhart carefully constructed within his own work.
5Second Cipher Solved
Building on this success, Fable 5.1 decrypted Urquhart's larger Cyphral Octastich from The Jewel (1652) - a 285-number cryptogram using the same principle but with pages as reference points instead of paragraphs. The decrypted eight-line royalist prayer in ottava rima form (ABABABCC) contained only nine unreadable letters out of 285 positions. Analysis showed Urquhart consistently used the first occurrence of needed letters (231 of 275 readable positions), with minor deviations explainable by transcription issues like hyphenated words and Greek phrases.
6Technical Validation
The team developed multiple verification tools including verify_octastick.py (mapping all 285 positions), jewel_pages.py (page parser), and analysis tools like firstocc.py. They confirmed 231 positions matched first-occurrence words exactly, with 44 minor deviations (1-3 words off) due to identifiable transcription factors. The solution required compensating for a single page numbering anomaly where position 159 onward needed a -1 page adjustment, demonstrating the AI's ability to identify and correct for historical text inconsistencies during decryption.
7Historical Cryptography Implications
This breakthrough suggests many historical ciphers may require contextual rather than mathematical solutions. Urquhart's approach - hiding messages through self-referential textual structures rather than complex encryption - represents an understudied cipher tradition. The success demonstrates how AI can surpass human cryptographers by making unconventional connections between a cipher and its host text's structure, literary devices, and historical context - capabilities that could revolutionize historical cryptography research.
8Future of AI in Historical Research
The dual cipher solution establishes AI as a powerful tool for humanities research, capable of processing historical texts at scale while recognizing subtle contextual patterns. This achievement suggests AI could help reconstruct lost or damaged historical documents, identify unknown authors through stylistic analysis, and solve other long-standing literary mysteries. As the team notes, full verification of the Octastich solution awaits examination of physical 1652 editions, highlighting how AI can direct traditional scholarly research toward promising avenues of investigation.