The Voynich manuscript has resisted every code-breaker for 600 years

Few objects have frustrated linguists, mathematicians, and professional cryptographers as thoroughly as a small illustrated codex written in a script that no one has ever been able to read. Bound in worn vellum and filled with drawings of plants that match nothing in the botanical record, it has become one of the most famous unsolved puzzles in the history of writing.

A book of vellum, plants, and stars

The volume runs to roughly 240 pages of animal-skin parchment, densely covered in a flowing, looping hand that reads left to right. Its pages are grouped into loose sections that scholars have nicknamed by their imagery: a botanical section crowded with drawings of flowers, roots, and leaves; an astronomical section full of suns, moons, and star-studded circles; a biological section showing tiny human figures bathing in green pools connected by tubes; and a pharmaceutical section that pairs labeled jars with plant parts.

What makes the drawings so unsettling is that almost none of the plants can be matched to real species, and the astronomical diagrams do not line up neatly with any known medieval system. The pictures look purposeful and carefully organized, yet they resist every attempt to tie them to a recognizable field of knowledge.

The layout adds to the mystery. Text wraps neatly around the illustrations as though written to accompany them, and labels appear beside individual plant parts, jars, and celestial figures, exactly as they would in a genuine reference book of the period. That deliberate structure is part of what makes the volume so hard to dismiss out of hand, because a work of pure nonsense rarely takes such consistent pains to look like an ordinary manual of medicine or natural philosophy.

Carbon dating points to the early 1400s

For decades the age of the manuscript was itself a matter of dispute. That question was largely settled when a sample of the parchment was subjected to radiocarbon analysis, which placed the animal skins in the first decades of the fifteenth century. That dating, combined with the style of the script and pigments, anchors the object firmly in the late medieval or early Renaissance period rather than in any later era.

The physical volume is preserved today at Yale University, where it is cataloged as Beinecke MS 408 and made available in high-resolution digital form. Its full record and page-by-page scans sit in the Beinecke Rare Book and Manuscript Library collection, which has allowed researchers around the world to study every stroke without touching the fragile original.

The script no one can read

The writing itself, often called Voynichese, behaves in ways that both invite and defeat analysis. It contains a limited set of recurring characters that combine into word-like units, and those units follow statistical patterns similar to real languages, including predictable word lengths and repetition. Yet no one has matched the symbols to any known alphabet, and no reliable translation of even a single sentence has ever been demonstrated.

Some sequences repeat several times in a row, a feature that is rare in natural writing, while certain characters appear almost only at the beginnings or ends of lines. Those quirks have fueled endless debate about whether the text encodes a real message, a constructed language, or something else entirely.

Codebreakers who came up empty

The manuscript takes its common name from Wilfrid Voynich, a book dealer who acquired it in the early twentieth century and circulated it among specialists in hopes of a solution. In the years that followed, some of the most accomplished cryptanalysts of the modern era took a turn at it, including experts who had broken military ciphers during the world wars.

Every one of those efforts ultimately failed to produce a verifiable reading. Amateur and academic claims of a breakthrough surface every few years, often announced with great confidence, but none has survived scrutiny by other researchers. A broad overview of the many failed attempts and competing theories is maintained in the reference material on the Voynich manuscript.

Hoax, cipher, or lost language?

Explanations for the text tend to fall into three broad camps. One holds that it is a genuine cipher, concealing a real language behind a system of substitution or transposition that has simply never been cracked. A second argues that it records an otherwise unknown or invented language, perhaps a private script devised by a single author. A third suggests the whole thing is an elaborate hoax, a book of meaningless but convincing gibberish produced to be sold as a work of hidden wisdom.

Each theory has to contend with awkward facts. A pure hoax would be hard to reconcile with the language-like statistics of the text, while a real cipher or language should, in principle, have yielded at least a partial reading after a century of concentrated effort.

Why it still draws scholars

The manuscript endures as a magnet for study precisely because it sits at the intersection of several disciplines and satisfies none of them. Historians of science examine its illustrations, linguists model its symbol patterns, and computer scientists run it through the same tools used on ancient and encrypted texts. Its survival across roughly six centuries, largely intact and now freely viewable online, guarantees that the next generation of researchers will inherit the same tantalizing problem.

Modern computing has changed the tools without changing the outcome. Statistical models that measure the frequency and arrangement of symbols can characterize the text in exquisite detail, confirming that it behaves in language-like ways, but characterizing a pattern is not the same as reading it. The manuscript has become a benchmark of sorts, a problem stubborn enough to test whatever new method a researcher brings to it.

Whether the answer turns out to be a clever code, a forgotten tongue, or an ingenious fraud, the object has already earned its place as a rare puzzle that has outlasted everyone who ever tried to solve it.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview