The Voynich manuscript is a hand-written book with pictures, written in a language that no one understands. The paper used to make it has been dated to the early 15th century (1404–1438). Experts believe it may have been created in Italy during the Italian Renaissance. Scholars still debate where it came from, who wrote it, and what it was meant for. No one has found a translation or enough context to prove or rule out any of the many ideas about it. Some think it might be a script for a real or made-up language, a secret code, or even a fake book, a reference guide, a made-up language, or a fictional story.
The first known owner was Georg Baresch, a 17th-century alchemist from Prague. The book is named after Wilfrid Voynich, a Polish bookseller who bought it in 1912. The manuscript has about 240 pages, but some pages are missing. The text is written from left to right, and some pages are folded and different sizes. Most pages have strange drawings and diagrams, including people, unknown plants, and symbols related to the stars. Since 1969, it has been kept in Yale University’s Beinecke Rare Book and Manuscript Library. In 2020, Yale made the entire manuscript available online.
Many professional and amateur codebreakers, including some from World War I and World War II, have studied the manuscript. Experts like Prescott Currier, William Friedman, Elizebeth Friedman, and John Tiltman were unable to decode it. No one has successfully translated the text, and none of the ideas about its meaning have been proven. The mystery of the manuscript has led to much discussion and research.
Description
The physical features of the manuscript have been studied by many researchers. It measures 23.5 by 16.2 by 5 cm (9.3 by 6.4 by 2.0 in) and has hundreds of vellum pages grouped into 18 quires. The total number of pages is about 240, but the exact number depends on how the unusual foldouts are counted. The quires are numbered from 1 to 20 in several places using numerals similar to those from the 15th century. The top right corner of each recto (right-hand) page is numbered from 1 to 116 using a style of numerals that appeared later. Gaps in the numbering suggest the manuscript may have had at least 272 pages in 20 quires, some of which were missing when Wilfrid Voynich acquired it in 1912. Evidence shows that many pages were rearranged over time, and the current order differs from the original.
In 2009, samples from the manuscript were tested at the University of Arizona using radiocarbon dating. Results showed the parchment dates between 1404 and 1438. In 2014, protein testing confirmed the parchment was made from calfskin. Multispectral analysis showed the parchment was not written on before the manuscript was created. The parchment quality is average, with common issues like holes and tears, but it was carefully prepared so the skin side is nearly identical to the flesh side. The parchment was made from "at least fourteen or fifteen entire calfskins."
Some pages are thicker than usual. The goatskin binding and covers were added later when the manuscript was owned by the Collegio Romano. Insect holes on the first and last pages suggest a wooden cover was used earlier. Discoloration on the edges indicates a tanned leather cover inside.
Medievalist Lisa Fagin Davis described the parchment as soft, like books that have been "heavily thumbed." This suggests the manuscript was used frequently, possibly as a medical manual or celestial almanac. Its heavy use indicates it served a practical purpose rather than a sacred or ceremonial one. Holes from scabs, wounds, or insect bites, and the lack of luxurious features like gold leaf, support this.
Many pages have drawings or charts colored with paint. Modern analysis using polarized light microscopy showed a quill pen and iron gall ink were used for text and outlines. The ink used for drawings, text, and page numbers has similar microscopic features. In 2009, energy-dispersive X-ray spectroscopy found the inks contain carbon, iron, sulfur, potassium, and calcium, with small amounts of copper and zinc. Lead was not found, but X-ray diffraction identified potassium lead oxide, potassium hydrogen sulfate, and syngenite in one sample. Similarities in the ink suggest the drawings and text were created at the same time.
Colored paint was applied crudely to ink-outlined figures, possibly later. The blue, white, red-brown, and green paints were analyzed using polarized light microscopy, X-ray diffraction, energy-dispersive X-ray spectroscopy, and scanning electron microscopy. The pigments used were inexpensive.
Computer scientist Jorge Stolfi noted that some text and drawings were modified with darker ink over earlier, fainter writing. Evidence of this appears on pages like f1r, f3v, f26v, f57v, f67r2, f71r, f72v1, f72v3, and f73r.
The manuscript has about 38,000 words, with 9,000 being unique. Every page contains text, mostly in an unknown language, though some have Latin script. The majority of the text is written in an unknown script, left to right, with characters made of one or two simple strokes. Some debate exists about whether certain characters are distinct, but a script of 20–25 characters would cover most of the text. A few rare characters appear only once or twice. There is no obvious punctuation.
Most text is written in a single column with a slightly uneven right margin and paragraph divisions, sometimes with stars in the left margin. Other text appears in charts or as labels for illustrations. The writing flows smoothly, suggesting the symbols were not encoded. No delays between characters are present, as would be expected in encoded text.
Only a few words are thought to be written in a known language. Various alphabets, like the Extensible Voynich Alphabet (EVA), have been created to help analyze the text. The first major system was developed in the 1940s by a team led by cryptographer William F. Friedman, who transcribed the manuscript onto IBM punch cards for machine readability.
The text contains over 170,000 characters, divided into about 35,000 groups, usually called words or word tokens (37,919). Of these, 8,114 are unique. The structure of the words follows patterns similar to natural languages, such as rules about character placement and repetition. The most common word appears roughly twice as often as the second-most-common word, following Zipf’s law. Word lengths and the ratio of unique to total words are similar to languages worldwide. Some words appear in predictable sequences, as if following grammar rules. Different sections of the manuscript, such as those with plant drawings or star charts, have distinct sets of frequently used words. The language differs from known languages in its spelling, with many repeated letter sequences.
Professor Gonzalo Rubio, an expert in ancient languages, noted that grammatical markers, like "s" or "d" in English, never appear in the middle of words in the Voynich manuscript. This is unusual for Indo-European, Hungarian, or Finnish languages.
History
Much of the book's early history is unknown, though the text and illustrations are clearly European in style. In 2009, researchers at the University of Arizona used a scientific method called radiocarbon dating to determine that the manuscript's vellum (animal skin used for writing) was made between 1404 and 1438. Additionally, McCrone Associates, a laboratory in Illinois, found that the paints used in the manuscript were typical of that time period in Europe. Some incorrect reports claimed that McCrone said the ink was added later, but their official report does not support this.
The first known owner was Georg Baresch, a 17th-century alchemist from Prague. Baresch was confused about the "Sphynx" (a mysterious book) that had been in his library for years. He learned that Athanasius Kircher, a Jesuit scholar from Rome, had published a Coptic (Egyptian) dictionary and claimed to understand Egyptian hieroglyphs. Baresch sent Kircher samples of the script twice, asking for help. A letter from Baresch to Kircher, dated 1639, is the earliest confirmed mention of the manuscript.
It is unclear whether Kircher answered Baresch’s request, but Kircher wanted to own the book, which Baresch refused to give up. After Baresch died, the manuscript went to his friend Jan Marek Marci, who later became rector of Charles University in Prague. A few years later, Marci sent the book to Kircher, along with a letter in Latin (dated August 19, 1665 or 1666) that was still attached to the book when Wilfrid Voynich later acquired it. The letter stated:
"Reverend and Distinguished Sir, Father in Christ:
This book, given to me by a close friend, I have always intended to send to you, my dear Athanasius, because I believed only you could read it. The previous owner once asked for your help by sending you a copy of the script, but he refused to give you the book itself. He worked hard to decipher it but gave up before he died. Now, I send you this gift, which I hope you will succeed in unlocking.
Dr. Raphael, a tutor to King Ferdinand III of Bohemia, told me the book once belonged to Emperor Rudolf and that Rudolf gave 600 ducats (a large amount of gold) to the person who brought it to him. He believed the author was Roger Bacon, an English scholar. I do not know if this is true, but I leave it to you to decide.
At the command of your Reverence,
Joannes Marcus Marci of Cronland
Prague, 19th August, 1665 [or 1666]"
The "Dr. Raphael" is believed to be Raphael Sobiehrd-Mnishovsky, and the 600 ducats equal about 2.10 kilograms of gold. Records from Rudolf’s time show a similar transaction in 1599, when Rudolf bought "a couple of rare books" from Carl Widemann for 600 florins. Widemann collected alchemical manuscripts, so it is possible he owned the book, though this is not proven.
Wilfrid Voynich believed Raphael’s claims, but the idea that Roger Bacon wrote the book has been largely rejected. However, the book’s first page has a faint name, Jacobus Horcicky de Tepenecz, who was head of Rudolf’s botanical gardens in Prague. Rudolf owed money to de Tepenecz, and it is possible de Tepenecz received the book as payment.
No records of the book’s ownership for the next 200 years have been found, but it likely stayed in the library of the Collegio Romano (now the Pontifical Gregorian University) in Rome. When Italian troops captured Rome in 1870, the new Italian government took Church property, including the library. Many books were moved to faculty libraries to avoid being taken, and the Voynich manuscript was among them. It still has the mark of Petrus Beckx, who was head of the Jesuit order and the university’s rector at the time.
Beckx’s private library was later moved to the Villa Mondragone near Rome, a palace owned by the Jesuits. In 1912, the Jesuits sold some manuscripts to the Vatican Library, but not all went there. Wilfrid Voynich bought 30 manuscripts, including the Voynich manuscript, in 1912. He spent seven years trying to interest scholars in studying it.
In 1930, the manuscript was inherited by Wilfrid’s wife, Ethel Voynich, who later gave it to her friend Anne Nill. Nill sold it in 1961 to Hans P. Kraus, an antique book dealer. Kraus could not find a buyer and donated the manuscript to Yale University in 1969, where it was cataloged as "MS 408" or "Beinecke MS 408."
Authorship hypotheses
Many people have been suggested as possible authors of the Voynich manuscript, including Roger Bacon, John Dee, Edward Kelley, Giovanni Fontana, and Voynich himself.
In a letter written in 1665 or 1666, Marci said that his friend Raphael Mnishovsky, who had died in 1644, claimed the book was once bought by Rudolf II, Holy Roman Emperor and King of Bohemia, for 600 ducats and 2.10 kilograms of gold. Mnishovsky believed the author might have been Roger Bacon, a 13th-century Franciscan friar and scholar. Marci did not confirm this claim, but Wilfrid Voynich took it seriously and considered the possibility that Albertus Magnus might have written it instead.
The idea that Bacon wrote the manuscript led Voynich to think John Dee might have sold it to Rudolf II. Dee was a mathematician and astrologer at Queen Elizabeth I’s court who owned many of Bacon’s writings. Dee and his assistant Edward Kelley lived in Bohemia for years, hoping to sell their services to the emperor. However, Dee’s detailed records do not mention this sale, making it unlikely. If Bacon did not write the manuscript, the connection to Dee becomes weaker. Before carbon dating, it was thought that Dee or Kelley might have written it and spread the rumor that Bacon created it to sell it later.
Some believe Voynich might have made the manuscript himself. As an antique book dealer, he had the knowledge and resources to do so. A lost book by Bacon would have been valuable. However, letters from Baresch and Marci only mention a manuscript, not the Voynich one specifically. These letters might have inspired Voynich to create the manuscript, but experts say the manuscript’s internal dating and the discovery of Baresch’s letter in 1999 make this unlikely.
Eamon Duffy said that the radiocarbon dating of the vellum proves the manuscript is not a fake from the post-medieval period. The pages are consistent, showing they came from a single source. It is impossible that so much unused parchment from the early 15th century could have survived.
Some think illustrations in Giovanni Fontana’s books resemble those in the Voynich manuscript. Fontana used cryptography in his works, but he used a simple substitution cipher, not the Voynich script. In one of his books, he described machines using his cipher.
Before 1921, Voynich saw a faint name at the bottom of the manuscript’s first page: “Jacobj à Tepenecz.” This is believed to refer to Jakub Hořčický of Tepenec, a man Rudolf II ennobled and employed as a physician and curator of his gardens. Voynich and others thought Jacobus owned the manuscript before Baresch, linking it to Rudolf’s court. However, Jacobus’s name has faded, and it does not match his signature in a document found in 2003. This has led some to think the name was added later, possibly by Voynich.
Baresch’s letter is similar to a trick played by Andreas Müller on Athanasius Kircher. Müller sent Kircher an unintelligible text from Egypt, asking for a translation. Kircher claimed to solve it, but some think Müller and Baresch might have tested Kircher’s skills.
Raphael Mnishovsky, the source of the Bacon story, was a cryptographer who claimed to have created an unbreakable cipher around 1618. This has led some to think he might have made the Voynich manuscript to test his cipher, with Baresch as an unknowing test subject. The disclaimer on the manuscript’s cover might suggest Marci suspected a trick.
In 2006, Nick Pelling proposed that Antonio Averlino, a 15th-century Italian architect, wrote the manuscript. This theory fits the radiocarbon dating. Jules Janick and Arthur O. Tucker, based on plant and animal identification and a map of Mexico, suggested the manuscript was created in Mexico between 1562 and 1572.
Language hypotheses
Many ideas have been proposed about the "language" in the Voynich manuscript, called Voynichese.
One idea is that the manuscript contains a meaningful text in a European language, but it was made hard to understand by using a secret code. This code changed each letter in the original text to a different symbol. This was the main idea for most 20th-century attempts to decode the manuscript, including a group of experts from the NSA led by William F. Friedman in the 1950s.
However, many experts say this idea is unlikely. For example, simple codes that replace each letter with another letter do not match the patterns in the manuscript. Also, the small number of symbols used suggests that other types of codes, like those with many symbols, are not the right fit. Codes called polyalphabetic ciphers, which were created in the 1460s, usually produce text where each symbol appears about the same number of times. But the Voynich manuscript seems to have patterns more like real languages.
Some patterns in the manuscript, like repeated shapes such as "or," "ar," and "an," suggest a different kind of code. This code might turn each letter in the original text into a group of symbols. For example, lines in the manuscript look similar to how Roman numerals might be written if they were encoded this way. A study from 2025 proposed a possible code called "Naibbe" that could explain the manuscript's patterns. However, the study says this is just an example and not proof that this is the actual code used.
In 1943, Joseph Martin Feely thought the manuscript might be a scientific diary written in a private shorthand system using abbreviations. However, other experts later disagreed with his ideas.
Another idea is that the text is mostly meaningless, but important information is hidden in small details, like the second letter of each word or the number of letters in each line. This method is called steganography and was described in 1499. Some people thought a tool called a Cardan grille might have been used to hide the message, but this is unlikely because the text does not appear to be arranged in a regular pattern.
Statistical analysis shows the manuscript has patterns similar to real languages. For example, the complexity of the words is similar to that of English or Latin. A linguist named Jacques Guy once thought the text might be a little-known natural language written with a made-up alphabet. He joked about it being Chinese, but later studies made him take the idea seriously. Other experts have suggested the text might be a previously unknown dialect of a Germanic language.
In 2014, Stephen Bax tried to decode the manuscript by looking for proper nouns, like names, and matching them to illustrations. He translated some symbols and words, suggesting the text is about nature written in a real language. However, no further work has been done since his death in 2017. Some researchers have used computer programs called deep learning to study the manuscript's symbols. In 2023, these programs found that the manuscript's symbols most closely match an ancient Indian script called Khojki.
In 2014, a study claimed the manuscript includes references to 37 plants, 6 animals, and one mineral found in an Aztec herbal book. The researchers also found a mineral called atacamite in the manuscript's paint, suggesting the text might be in Nahuatl, the language of the Aztecs. They dated the manuscript to between 1521 and 1576. However, this date conflicts with earlier tests of the manuscript's materials, which suggested it was older. Some experts argue that a skilled forger could have created the plants to look like real ones. A Nahuatl expert named M. Pharao Hansen said the study's conclusions were based on "pure speculation."
William F. Friedman thought the text might be a made-up language. In 1950, he asked John Tiltman, a British officer, to analyze the text, but Tiltman did not agree. Tiltman later wrote that the text seemed to mix different types of codes.
The idea of a made-up language is not new. In 1668, a man named Bishop Wilkins created a system to classify ideas using symbols. However, this was much later than when the Voynich manuscript is believed to have been written. In many made-up languages, words are formed by adding parts to a base word, which could explain the repetition of certain letters in the manuscript.
Decipherment claims
Since the manuscript was found again in 1912, many people have tried to decode its writing.
One of the first attempts was made in 1921 by William Romaine Newbold from the University of Pennsylvania. He believed the visible text had no real meaning, but each letter was actually made up of tiny marks only visible under a magnifying glass. He claimed these marks were based on ancient Greek shorthand, forming a second layer of writing that contained the real message. Newbold said he used this idea to translate entire paragraphs, proving the author was Roger Bacon and that Bacon used a microscope 400 years before van Leeuwenhoek. A drawing in the astronomy section showed an object with four curved arms, which Newbold thought was a galaxy, something only visible with a telescope.
Later, John Matthews Manly from the University of Chicago pointed out serious problems with Newbold’s theory in 1931. For example, the shorthand characters could have many meanings, making it impossible to know which was correct. Newbold’s method also required rearranging letters until they formed readable Latin, which made the system too flexible to prove anything for sure. Studies showed the markings were actually caused by ink cracking on rough vellum, and seeing meaning in them was likely due to pareidolia, or seeing patterns where there are none. Because of Manly’s work, the idea of micrography is no longer widely accepted.
In 1943, Joseph Martin Feely published a book claiming the manuscript was a scientific diary written by Roger Bacon. He said the text was a shortened form of medieval Latin using a simple substitution cipher.
Leonell C. Strong, a scientist and amateur cryptographer, proposed in 1947 that the manuscript used a complex system of numbers and letters. He translated two pages, claiming the text was written by Anthony Ascham, a 16th-century writer. However, notes after his death showed his method was subjective and unreliable.
In 1978, Robert Brumbaugh, a professor at Yale University, said the manuscript was a fake meant to trick Emperor Rudolf II. He believed the text was Latin encoded with a two-step method.
Also in 1978, John Stojko claimed the manuscript was written in vowelless Ukrainian. His theory gained attention in Ukraine but was criticized for lacking evidence and having a weak method.
In 2014, Stephen Bax, a professor of linguistics, proposed a partial translation of the manuscript using methods similar to those used for Egyptian hieroglyphs. He suggested the text was a treatise on nature in a Near Eastern or Asian language, but no full translation was completed before his death in 2017.
In 2017, Greg Kondrak, a professor at the University of Alberta, and his student Bradley Hauer used computer analysis to suggest the manuscript might be written in Hebrew, encoded as anagrams. However, experts in medieval manuscripts were not convinced by their findings.
In 2017, Nicholas Gibbs, a television writer, claimed the manuscript was written in abbreviated Latin and was mostly copied from existing works about women’s health. Scholars criticized his translation as incorrect and not meaningful.
In 2018, Ahmet Ardıç, a Turkish Canadian engineer, claimed the manuscript was written in Old Turkic, using phonetic spelling. He said he found similarities between the text and Turkish words. His team translated about 300 words, but their submission to a journal was rejected in 2019. Critics compared his work to discredited theories like the Sun Language Theory.
In 2019, Gerard Cheshire, a biology researcher, claimed the manuscript was written in a "calligraphic proto-Romance" language. He said it was a guide to herbal remedies, health, and astrology, written by Dominican nuns for Queen Maria of Castile. He later translated an illustration on page 158, claiming it showed a volcano near Vulcano Island. However, experts disputed this interpretation.
Facsimiles
Many books and articles have been written about the manuscript. In 1637, alchemist Georg Baresch made copies of the manuscript pages and sent them to Athanasius Kircher. Later, Wilfrid Voynich also created copies. In 2004, the Beinecke Rare Book and Manuscript Library made high-resolution digital scans of the manuscript available online for the public. Printed copies of the manuscript also became available at this time. In 2016, the Beinecke Library and Yale University Press worked together to publish a detailed copy of the manuscript, called The Voynich Manuscript, along with scholarly essays. The Beinecke Library also allowed a Spanish publisher named Siloé to print 898 copies of the manuscript in 2017. In September 2024, special scans of ten pages were made public, showing details that cannot be seen with normal light.