A new peer-reviewed study proposes that generations of codebreakers may have been trying to read the Voynich Manuscript in an order its makers never intended. The famously undeciphered medieval book, now held by Yale’s Beinecke Library, may originally have been a stack of individual folded parchment sheets that were later nested and bound into the wrong sequence.
That claim is more than a fresh coat of mystery paint. Researchers Colin Layfield and Lisa Fagin Davis point to several kinds of evidence that seem to agree with one another: scribal hands alternate strangely in the present binding, some illustrations appear to continue across pages now separated by other leaves, and a statistical comparison of the text finds unlikely neighbors alongside pages that look as though they belong together. The proposed answer is not “we cracked Voynichese.” It is that scholars may first need to reconstruct the book before they can fairly judge the writing inside it.
The new claim is about the book, not the code
The study, “Singulion Structure and the Voynich Manuscript,” was published in Digital Medievalist. It argues that the manuscript may have been designed as a sequence of “singulions”: separate sheets of parchment, each folded once to create four pages, stacked one after another.
That differs from the more familiar medieval method of book construction. Scribes commonly nested several folded sheets to form a gathering, or quire, then bound multiple gatherings together. If a later binder mistook a stack of separate folded sheets for material meant to be nested, the resulting book could preserve every sheet while scrambling the intended page order. It would be the codex equivalent of shuffling a deck, then stapling it shut.
The physical artifact matters here. Yale identifies the volume as Beinecke MS 408, a manuscript written in an unidentified script and filled with plant-like drawings, astronomical diagrams, human figures and other images that have resisted a settled interpretation. Its parchment has been radiocarbon-dated to the early 1400s, but its authors, language and purpose remain unknown.
The archival Yale lecture below introduces the manuscript as a physical object and features Davis discussing what its pages can reveal. It predates the 2026 singulion study and should not be mistaken for footage of the new research itself.
Five scribes should not trade places like this
Davis’s earlier handwriting research identified five different scribal hands in the manuscript. In an ordinary production, a scribe might copy a coherent run of leaves before another person takes over. The Voynich sequence instead makes the hands appear to alternate in ways that are difficult to explain as a normal division of labor.
The singulion model offers a practical explanation. If each contributor wrote on separate folded sheets, those sheets could later have been stacked or circulated before binding. Nesting them incorrectly would make the same writer disappear and reappear through the finished book, even if the original packets had been orderly.
This does not prove who the scribes were or whether they were writing a natural language, a cipher, a constructed system or something else. It does make the manuscript look less like one eccentric author improvising hundreds of pages and more like a shared project whose physical history matters as much as its symbols.
The pictures seem to cross gaps that may not belong there
Some visual clues are wonderfully stubborn. Davis told Live Science in its October 20 report that certain images run off one page and appear to resume on another even though multiple pages intervene in the current order. A misbound set of folded sheets could bring those separated halves back into logical contact.
The manuscript also has a large water stain along the upper margins of some leaves. The researchers plan to compare the stain’s shapes because liquid passing through a stacked manuscript should leave a physical sequence. If adjacent stain patterns no longer align, they could become a kind of accidental map for reconstructing the earlier order.
That work is still proposed, not completed. It is one of the study’s strengths that the claim produces a testable next step instead of demanding faith in a dramatic translation.
The text has neighbors—and apparent strangers
Layfield also used latent semantic analysis, a statistical method for comparing patterns within blocks of text. The technique cannot translate Voynichese when nobody knows what the symbols mean. It can, however, ask whether two passages behave as though they discuss related material.
According to the researchers, some pages that sit next to each other now have unexpectedly low similarity, while pages farther apart show stronger relationships. That pattern does not independently prove the singulion theory, but it supports the physical and visual clues rather than fighting them. The shuffled-book idea becomes interesting precisely because handwriting, illustration and computation all point toward the same structural problem.
What this changes—and what remains unsolved
The study does not reveal a dictionary, identify the strange plants or translate a single recipe. Any headline claiming that the Voynich Manuscript has finally been solved would be doing the mystery a disservice.
What it changes is the starting assumption. Statistical studies, word-position models and attempts to trace narrative flow often treat the present page order as meaningful. If the binding scrambled that order centuries ago, some apparent discontinuities may belong to the book’s construction rather than its language. Previous analyses do not become worthless, but sequence-dependent results may need to be tested again against a reconstructed arrangement.
There is also a deliciously nerdy reversal here. The obstacle may not be an impossibly clever code alone. It may be the oldest troubleshooting lesson in the manual: before blaming the software, check the hardware. The Voynich Manuscript is still unreadable, but researchers now have a concrete way to ask whether everyone has been opening the right pages in the wrong order.
