No one has ever cracked the script filling a medieval book of impossible plants

Inside a climate-controlled vault at Yale University sits a small vellum codex covered in flowing, looping script that resembles no known language on Earth. Its pages are filled with drawings of plants that do not match anything in nature, star charts that follow no known system, and diagrams of tiny nude figures bathing in green liquid, all wrapped around text that cryptographers, linguists and now artificial-intelligence models have spent more than a century failing to translate.

Formally cataloged as Beinecke MS 408 but universally known as the Voynich manuscript, the book was produced in Central Europe toward the end of the 15th century or during the 16th, based on scientific testing of its materials. What makes it a standing mystery is not just that the writing is unread — it is that no other surviving document anywhere uses the same script, leaving researchers with no comparison point to even begin cracking it.

An Alphabet That Exists Nowhere Else

The script, nicknamed “Voynichese,” follows internally consistent patterns that resemble real language: letter frequencies, word-length distributions and repeated structures that look statistically like natural writing rather than random marks. Yet no proposed match to a known language, dialect, or historical cipher system has held up under scrutiny, and cryptographic approaches applied over decades have failed to produce a consistent decoding, according to Yale’s Beinecke Library, which has held the manuscript since 1969. Chemical analysis of its inks and pigments points to materials consistent with the period, reinforcing that the document is a genuine medieval artifact rather than a later forgery.

Six Sections, No Translation

Although the text resists reading, the manuscript’s roughly 240 surviving vellum pages are organized into six distinct visual sections that hint at its intended subject matter. One presents drawings of plant specimens unlike any identifiable species. Another contains astronomical and astrological diagrams. A third, the so-called balneological section, depicts small human figures bathing in interconnected tubs and pipes of green fluid. There are cosmological medallions, a pharmaceutical section illustrating herbs and roots alongside labeled jars, and a final block of unillustrated text, possibly recipes, in which each entry is marked by a small star-shaped flower in the margin.

A Well-Documented Journey Through Royal and Jesuit Hands

Unlike many mysterious artifacts, the manuscript’s ownership trail after 1600 is relatively well established. Its first confirmed owner was Holy Roman Emperor Rudolf II, who reigned from 1576 to 1612 and reportedly paid 600 gold ducats for it, apparently believing it to be a lost work of the medieval English friar Roger Bacon — a claim modern scholars consider pure speculation. The book later passed through the library of pharmacist Jacobus Horčicky de Tepenec and, by 1637, alchemist Georgius Barschius, before reaching Jesuit scholar Athanasius Kircher in 1666. It remained largely out of public view until 1912, when Polish-American book dealer Wilfrid Voynich purchased it from the Jesuit College at Frascati near Rome, giving the manuscript the name it carries today. After stops with Voynich’s wife Ethel and their assistant Anne Nill, New York bookseller H.P. Kraus acquired it in 1961 and donated it to Yale’s Beinecke Library in 1969.

Radiocarbon Dating Narrowed the Window, Not the Meaning

Separate from ink and pigment analysis, radiocarbon testing of the vellum itself has helped pin down when the pages were made. Results place the parchment’s production in the early 15th century, and Wikipedia’s summary of the manuscript’s research history notes that this dating range is one of the few points on which most specialists agree, even as they disagree about nearly everything else concerning the book. Establishing the vellum’s age ruled out several early theories, including the idea that the entire manuscript was a 19th- or 20th-century fabrication created to fool a collector, since the material itself predates any plausible forger by hundreds of years.

Theories Range From Language to Hoax

The absence of a confirmed reading has left the field wide open to competing explanations. Some researchers argue the text encodes a genuine natural language written in an invented alphabet; others propose it is an artificial or constructed language designed for the purpose; still others treat it as an unbroken cipher layered over ordinary European text. A minority view holds that the entire document is an elaborate hoax, whether a medieval con or, less plausibly given the confirmed period materials, a more recent fabrication. Modern computational linguists and, more recently, large language models have been applied to the puzzle, with a study published in early 2026 proposing the text may represent a specific type of medieval cipher — one addition to a long list of hypotheses that so far has not produced a scholarly consensus. Each new attempt tends to follow the same pattern: an initial claim of a breakthrough, followed by other researchers testing the method against the full text and finding it falls apart past the first few pages.

Preserved but Rarely Handled

Given the manuscript’s fragility and the volume of research requests it draws, Yale restricts physical access to the original and instead directs scholars and the public to high-resolution digital scans of every page. The book has left the Beinecke building only once since 1969, for a 2014-2015 exhibition on historical codes and ciphers at the Folger Shakespeare Library. Full facsimile editions have since been published for researchers who want to study the text and illustrations without handling the original vellum. Whether the next attempt to read it comes from a cryptographer, a historian or an algorithm, the manuscript’s basic challenge has not changed since 1912: a script with no known relatives, attached to a book whose author never left a name.

This article was produced with the assistance of AI and reviewed by an editor.

More from Morning Overview