The development of AI has brought promising results in decoding the mysterious 600-year-old manuscript, but are we really getting there?

The Voynich Manuscript, a mysterious tome discovered in 1912 by the Polish book dealer Wilfrid Voynich, has been a source of fascination and intrigue for over a century. Its pages are filled with an unknown script, bizarre illustrations, and no discernible author or origin, making it a prime subject for cryptologists and historians.

The manuscript is believed to have been created in the early 15th century, based on carbon dating of the vellum on which it is written. It is divided into several sections, each featuring its own set of illustrations and text, including herbal, astronomical, biological, cosmological, and pharmaceutical components. The illustrations range from unknown plants to zodiac symbols, and from naked women to strange objects, none of which have been conclusively identified or explained.

The surviving 240 pages of this mysterious manuscript hint that it might have been a guide for medieval medicine or a pharmacopoeia. The first part seems to be about plants and their uses, but it’s hard to figure out exactly which plants are shown (some, like the wild pansy and maidenhair fern, are recognizable though). The drawings of medicines look like they were copied, but with some strange details added in, while some of the plant pictures seem like they’re made up of parts from different plants. Astrology was probably important in the medicine of the time, as shown by Zodiac symbols and maybe even pictures of planets. And a section seems to be related to balneology, with illustrations mostly showing small ɴuᴅᴇ women who look like nymphs, some wearing crowns, bathing in pools or tubs linked by an intricate system of pipes.

But even with all this info, we’re still not sure what the manuscript was really meant for. It’s like a puzzle that gives us a peek into how people practiced medicine and used astrology back in medieval times.

Numerous decoding attempts have been made over the years by a wide range of scholars from historical codebreakers to modern computer scientists. The manuscript has even attracted the attention of famous cryptologists such as Alan Turing and the Bletchley Park team, though it has resisted all efforts to reveal its secrets. Theories about its content have ranged from it being an elaborate hoax to the work of aliens, but none have been proven.

In recent years, artificial intelligence has joined the quest to decode the Voynich Manuscript. Most notably, a few years ago University of Alberta researchers Greg Kondrak and Bradley Hauer employed AI-based algorithmic decipherment techniques to finally crack the enigma. And did the results seem promising… Initially, the data suggested the text is in Arabic, but further analysis suggested Hebrew with a surprising 97% accuracy based on letter frequency.

The team then theorized the text might be an alphagram, where letters are shuffled and lack vowels. Unscrambling the first line of the manuscript using this method, they found 80% of the resulting words existed in a Hebrew dictionary. However, a native Hebrew speaker found the “deciphered” sentence nonsensical. Even with spelling corrections and machine translation (well, Google Translate, to be precise), the final result, “She made recommendations to the priest, man of the house and me and people,” remained grammatically correct but semantically unclear.

The “translation” of another 72-word section included the words “farmer,” “light,” “air” and “fire,” but since the translations produced did not form coherent sentences in Hebrew, the findings have been met with skepticism from the academic community.

But Kondrak and Hauer themselves acknowledge their research is preliminary, and it was indeed the press that, upon seeing the results, came up with headlines that were too sensational and exaggerative. In fact, many experts do recognize the value of their underlying algorithms, at the same time pointing out that crucial steps were overlooked before any ᴀssertions about the manuscript itself should have been made.

So for now, the mystery lives on, and the Voynich Manuscript remains largely undeciphered. But new contenders keep emerging one after the other, each ᴀsserting they’ve unraveled the mystery. A recent example is that of Gerard Cheshire, a University of Bristol biology research ᴀssistant, who gained attention for proposing the Voynich manuscript was written in “calligraphic proto-Romance,” making it the only document written in that language. He claimed to have decoded the text in two weeks, suggesting it’s a compendium on herbal remedies, astrology, and women’s health, compiled by Dominican nuns for Queen Maria of Castile.

However, medieval document experts, including Lisa Fagin Davis, criticized Cheshire’s method as circular and aspirational, dismissing his translations as gibberish and his claims as unfounded (with the University of Bristol distancing itself from the study shortly thereafter). That said, the proposition that the book is about women’s health is interesting — at least, in medieval times this could indeed have been a subject so mysterious and off-limits that it required being hidden away in undecipherable text.

Well, hopefully one day we’ll see. Kondrak and Hauer’s algorithm, and the use of artificial intelligence in general, definitely holds promise in terms of solving the mystery of the Voynich Manuscript. Of course, this might not be such good news for experts who have spent years and years of their lives trying to decode the mysterious document, constantly holding onto hope that they’re making progress. But now, will some bloodless machine get there first instead, in the blink of an eye? That’s another mystery only time will crack.