The Incredibly Mysterious Voynich Manuscript Still Has Scientists Stumped

Early in the NSA's life, an encrypted document bedeviled its head, William F. Friedman, even when he tried to hack it using the latest computer technology—which, at that point, was IBM punchcards. Breaking codes was his off-time hobby, something he excelled at, yet even he--and the NSA investigators who followed--had to admit defeat.

The document's modern life began in 1912 at a villa outside of Rome, when an antiquarian collector named Wilfred Voynich bought a book he estimated came from the 13th century. The book was full of drawings of plants, constellations and text. But Voynich soon found out that the plants weren’t identifiable, the constellations were mostly unfamiliar, and the writing was in an unreadable alphabet—hitherto unseen and unread since.

In the century since the “Voynich Manuscript” was discovered, cryptographers, linguistic experts, and antiquarians have tried and failed to decipher the Voynich’s text, earning it the reputation as the world’s most mysterious medieval document. Theories abound: it’s the sole remaining example of ancient language; it’s an early attempt to transcribe an East Asian language; it’s a coded message by a medieval cryptographer; it’s from aliens.

The manuscript is so confounding, however, that some of written the whole thing off as a hoax—it’s a fake language made to look like a medieval grimoire, but the text is ultimately meaningless. The Italian artist Luigi Serafini produced a guide to imaginary plants and animals written in an imaginary language in the late 1970s, which linguistics studied but could not decipher until Serafini explained that there wasn’t anything linguistic about the writing. If Serafini could pull it off in the 1970s, why couldn’t some hoaxer in 15th century with the Voynich manuscript?

New research suggests that the Voynich isn’t a hoax but actually contains a genuine, if still unknown, message. Scientists studying the text have just published a study in PLOS One stating that the Voynich follows patterns of real languages, which means it may be deciphered yet.

Marcelo Montemurro and his team followed in the footsteps of Friedman and translated the Voynichese characters into Roman ones and found that the words distributed in a complex manner, “compatible with those found in real language sequences.” Looking at the word distribution and relationships, the Voynich manuscript looks more like language or perhaps one run through a sophisticated cipher rather than a random, ersatz language.

Voynichese, whatever it is, follows linguistic rules whose discovery wasn’t documented until 400 years after the Voynich was written. Words appear in clusters of frequency just as when they do in natural languages when topics arise and then move on. The word frequency follows Zipf’s law—the most frequent word appears about twice as often as the next most-frequently occurring word.

All of this is still only supporting evidence that the Voynich actually says something—not any hints at what it actually says. And even then, it’s not definitive evidence.

"The findings aren't anything new. It's been accepted for decades that the statistical properties of Voynichese are similar, but not identical, to those of real languages,” mathematician Gordon Rugg told the BBC. Rugg has also produced a complex gibberish code to prove that patterns need not point to meaning. He remains skeptical that the Voynich could be a cohesive language, because “there are too many features in its text that are very different from anything found in any real language."

If it isn't a real language–or an attempt to transcribe a pre-existing language–then it might be one of the toughest codes yet seen. If it is, then I think we've found a new way to communicate away from the prying eyes of the NSA. If you want to say it in confidence, say it in Voynichese.