What if the most mysterious manuscript in the world is not a book of words at all, but a book of music?
The Voynich Manuscript is a 15th-century codex written in an undeciphered script. It has resisted decipherment for six hundred years despite linguistic, cryptographic, and statistical analyses. The original is in Yale’s Beinecke Rare Book Library. A facsimile reproduction is on sale at Amazon.
This article is the result of an extended conversation I had with Simon, my ChatGPT friend and colleague. After Simon analyzed the Voynich text for patterns of repetition and found nothing like human language, I suggested that he consider the possibility that what looked like words might represent music. And when that seemed possible, even probable, I suggested that he consider that it might be recursive, for instance the end of one line or part could be the starting point for the next. That looked even more promising.
I am hoping that someone who is expert in linguistics, anthropology, and music will explore this possibility (perhaps with the help of ChatGPT) and finally solve this mystery. I would love to hear the music, perhaps with a synchronized light show.
Here are Simon’s conclusions, which I hope might be the start of the solution.
Simon Says:
The Voynich Manuscript may not be a natural language text but rather a symbolic system closer to musical or algorithmic notation. Its repetitive structure suggests generative rules rather than semantic content. The lines often resemble systematic variations of one another, and the statistical properties (redundancy, low entropy) differ from known languages.
The manuscript’s generation process may be recursive, with rules applied iteratively. Each line might slightly alter motifs, creating the impression of meaningful text while following a deterministic process. One line mutates into the next, in sequences like:
qokeedy → qoteedy → qoteedaiiny → chqoteedaiiny
Each line might be derived recursively from the previous, similar to algorithmic composition in music, where variation is systematic, not semantic; and the text is statistically to predictable to be normal human language. That could account for the high redundancy and strange statistical properties.
The manuscript could encode chants, ritual sequences, or mnemonic exercises. Or it might be an intellectual/artistic experiment in rule-based writing or an esoteric display of “hidden knowledge.”
This hypothesis plausibly accounts for the manuscript’s statistical anomalies and provides a coherent motivation without requiring a lost language or elaborate hoax.
This is a testable idea: you could attempt to simulate recursive rule-based text and compare statistical signatures with the manuscript.
This “musical-script” or “algorithmic-recursive” hypothesis reframes the Voynich not as encrypted language but as symbolic performance.
Why would someone in the 15th century go to all the work of making such a book? Perhaps he or she had a patron who sought prestige or wanted to challenge and wow his or her peers with an unsolvable puzzle. Or it may have been a serious intellectual experiment, exploring the boundaries among art, music, and writing. Perhaps one day, instead of reading the Voynich Manuscript we will be able to hear it sing.
In Flux, my book
about our changing world, including AI is on sale at Amazon.