The Voynich Manuscript: The Book Nobody Can Read
Klaus Schmeh
Skeptical Inquirer Volume 35.1, January/February 2011
For almost 100 years, experts and amateur researchers have tried to solve the riddle of a handwritten book, referred to as the “Voynich manuscript,” composed in an unknown script. The numerous theories about this remarkable document are contradictory and range from plausible to adventurous.
The facts regarding the Voynich manuscript can be told quickly. It is a handwritten book of 246 pages containing numerous illustrations and approximately 170,000 characters. What is special about it? The script employed is utterly unknown and therefore illegible. According to a radiocarbon analysis conducted in 2009 by the University of Arizona, the manuscript was created in the first half of the fifteenth century (probably between 1404 and 1438). So far, there is no written publication on this analysis, but one of the scientists involved in the examination confirmed by e-mail that a paper is scheduled for 2011.
The modern history of the Voynich manuscript began in 1912. At that time, a bookseller and book collector named Wilfried Voynich found it in an Italian Jesuit college. Further information is provided in a letter dated 1666, which-according to Voynich-was enclosed with the manuscript. This document names some other previous owners who had all lived in the first half of the seventeenth century, thus indicating that the manuscript had been written before then. On the basis of this letter, Voynich favored the English monk and Renaissance man Roger Bacon (1214–1294) as the book's author. However, this theory is now considered very improbable.
Not many more historic facts are known about the Voynich manuscript (Kennedy and Churchill 2005). In particular, it is unclear who wrote the book, what it contains, and what its purpose was. In light of the meager evidence, I-as a skeptic and member of GWUP (the German counterpart of CSI)-am not surprised at the great number of speculative theories about the mysterious script. I will present the most important ideas here.
A good point of entry into a Voynich analysis is certainly the script of the document itself. The author of the manuscript wrote from left to right-this can be discerned from the left-aligned formatting. The typeface and size of the characters are inconspicuous, which is not altered by the fact that the text contains no punctuation marks, because this is unexceptional for old texts. Thus it is evident to a layman, even before inspection of the illustrations, that the Voynich manuscript has its origins in European culture. Moreover, it is apparent that the author was quite accurate: there are no visible corrections in the text. Unfortunately, the Voynich text itself is not divided into chapters; there are no subheadings.
Uncommon Illustrations
Approximately 220 of the 246 Voynich pages are illustrated. Some of the pages can be unfolded, revealing illustrations that extend to several page lengths. Because, unlike the text, the illustrations can be divided into different sections, six chapters of the Voynich manuscript can be distinguished: the botanical chapter (with large plant illustrations), the astronomical chapter (with charts containing celestial bodies and the zodiac signs), the balneological chapter (with nude female figures in tubs), the cosmological chapter (with circles and rosettes), the pharmaceutical chapter (with plants, parts of plants, and pots), as well as a chapter with food recipes (without illustrations).
The different illustrations can hardly be related to a common topic. Thus the Voynich manuscript-if it has meaningful content at all-must be a treatise on many different subjects. One may possibly say that it is a textbook for magicians, physicians, pharmacists, and astrologers (when it comes to these professions, their borders were still blurred 500 years ago). Provided that it is hardly possible to recognize significant symbols and religious motives within it, the Voynich manuscript can neither be assigned to a certain school of thought nor to a particular religion.
Unfortunately, none of the 126 plant illustrations can be definitively identified. However, the plant pictures at least enabled certain conclusions regarding the date of origin, before the radiocarbon dating was performed. Comparisons of artistic styles showed that the manuscript presumably did not originate before the fourteenth century, which was, of course, later confirmed. Not confirmed, however, was a theory stated by the botanist Hugh O'Neill (O'Neill 1944). He considered two plant illustrations as representing sunflowers and identified another one as capsicum. Because both plants spread in Europe only after the “discovery” of America, their identification appeared to narrow down the period of origin. However, the two identifications O'Neill made are not precisely compelling, and thus O'Neill's conclusion-like so many others in connection with the manuscript-is just speculation.
It is hardly more illuminating to take a look at the astronomical and the cosmological sections, which contain pictures that can be identified as the Zodiac signs still familiar today (Aries, Taurus, Libra, and so forth). Scarcely another illustration in the Voynich manuscript is as unambiguous. Unfortunately, this observation does not result in further insight into the book's origin. The celestial bodies illustrated in the astronomical section cannot be identified and probably are only figments of imagination. Some Voynich researchers believe they recognize in these pictures the Andromeda fog or the Pleiades, but this again is just speculation.
The hairstyles and clothing of the people pictured in the book, as well as the style of the illustrations, were usually dated to the period 1450–1520, which proved reasonably compatible with the radiocarbon dating (between 1404 and 1438). In most cases, the pictured persons are naked women in big tubs filled with water, which makes conclusive interpretation of these illustrations in the context of fashion impossible.
Taking all facts into account, it is astonishing how little the numerous illustrations reveal about the Voynich manuscript. Does this make an argument for the whole document being meaningless? Or did the author intentionally choose ambiguous illustrations to prevent inferences about the encrypted (and therefore secret) text? I consider both explanations to be possible.
Cryptological Analyses
A glance at the pictures in the manuscript is indeed interesting, but as a cryptologist I am naturally more interested in the Voynich text. It is unclear whether it is an encrypted message or simply a text composed of unknown letters. This is irrelevant for cryptological analysis, because the use of unknown letters is also a form of encryption. For the analysis of an encrypted text, cryptology provides quite a number of statistical methods-for example the determination of letter frequencies. Some of these analyses indicate that the Voynich manuscript is composed in a usual language but written in unknown letters. There are between fifteen and twenty-five different letters in the manuscript, but in many cases it is not clear whether identical or different symbols have been used. For this same reason, letter frequency cannot be determined clearly. Nevertheless, the language of the manuscript can be brought in line with European languages, because the average word length is four or five letters. Following this line of consideration, arguments can be put forward that Greek, Latin, or one of several other European languages was used to compose the Voynich manuscript. It is a pity that this approach does not implicate a specific language.
However, the language of the manuscript does not correspond to any European language because the Voynich has no two-letter words or words with more than ten characters. Moreover, it is curious that some words are repeated successively up to five times. The distribution of the letters within each word also does not answer known language patterns. Looking at the text as a whole, far fewer recurring words turn up than would be expected. Such arguments reveal with a high probability that-against all appearance to the contrary-we are not dealing with a simple substitution of letters. There also is no clear evidence that other simple encryption methods were used.
A study by the philosopher William Newbold took another direction. Newbold declared he had solved the Voynich mystery in 1921 (Newbold 1928). He considered as relevant not the letters themselves but the small, barely visible marks applied to them. These marks supposedly formed Greek characters, making up a text that could be decoded into a meaningful message. The result seemed to be sensational: the produced message not only confirmed Roger Bacon as the manuscript's author, but it reputedly also revealed that Bacon already had a telescope at his disposal and knew the spiral structure of the Andromeda galaxy-either of which would revolutionize the history of science. But as expected, Newbold's decryption came across as largely arbitrary and moreover only worked for a short section of the text. Therefore Newbold's theory could not prevail.
In 1943, the lawyer Joseph Feely published a cryptological paper regarding the Voynich manuscript. Feely, too, presented as a direct result of his research the supposed solution of the Voynich encryption (Feely 1943). By means of statistical analyses, he had come to the conviction that the manuscript was composed in Latin and contained numerous abbreviations and abbreviated sentences. With this basis, Feely translated a forty-one-line section of the manuscript. Unfortunately, Feely's approach made no sense at all and therefore quickly turned out to be a further dead end in Voynich research.
The U.S. cryptologist William Friedman (1891–1969) was considerably more competent. He is regarded as the most successful code-cracker of all ages; his name guarantees cryptological quality. In the course of his forty-year career, Friedman examined thousands of encryption methods during his service for the U.S. military and solved almost all of them. Unfortunately, despite all the texts he successfully analyzed, he had to surrender in the case of the Voynich manuscript. He therefore could not bequeath more to posterity than an educated guess. Friedman considered the text to be a treatise composed in an artificial language.
More Current Studies
The next Voynich studies worth mentioning originate with Robert Brumbaugh, a professor of the philosophy of the Middle Ages. He holds that the unknown characters are numerals that are each assigned several letters of the Latin alphabet (Brumbaugh 1978). However, the decryption provided by Brumbaugh did not make any sense. Another dispensable Voynich analysis was published by the physician Leo Levitov in 1987 (Levitov 1987). He, too, believed that he had decrypted the text. According to Levitov, the book is composed in an old form of Flemish that assimilated German and French words. Levitov supposed that in this manner a literary language emerged that served as an alternative to the Latin language common at that time. According to Levitov, the manuscript turned out to be written by the Cathari in the Middle Ages. However, Levitov's paper is so full of speculative assumptions that it can barely be taken seriously.
The British linguist Gordon Rugg is among the most reputable Voynich researchers. He conducted a most interesting cryptological experiment. For his experiment, Rugg generated a table with random combinations of characters that he used as prefixes, roots, or suffixes of new words. He positioned a quadratic stencil, like the ones used for encryption in the sixteenth century, over the table. In this manner he obtained a sequence of letters that bore great resemblance to the text of the Voynich manuscript. Rugg's experiment supports the hypothesis that the manuscript is nothing but a compilation of meaningless lines of letters (the hoax hypothesis) (Rugg 2004). The hoax hypothesis is backed by a text analysis by the Austrian physicist Andreas Schinner. Schinner discovered unnatural regularities in the word order of the manuscript that do not occur in any known language. He therefore also came to the conclusion that the Voynich manuscript is a fraud's artful fabrication, containing merely meaningless nonsense (Schinner 2007).
A relatively new theory was published by Briton Nick Pelling (2006). He considers the Italian architect Antonio Averlino (1400–1469) to be the Voynich author. Pelling supposes that Averlino escaped to Constantinople (Istanbul) around the year 1465, having beforehand recorded his knowledge as encrypted in the Voynich manuscript. Pelling provides numerous cryptological analyses that supposedly allow us to infer an applied method, but he does not present a solution. If we accept an inaccuracy of a few decades, Pelling's theory is consistent with the radiocarbon analysis. However, I consider the Averlino hypothesis very speculative. In addition, it seems improbable that the Voynich manuscript-which without any doubt would have attracted attention in the course of a luggage inspection-merely served as a means of secure transport.
What Is Behind It?
Looking at the diverse cryptological analyses of the Voynich text, we come to a similar result as we did regarding the illustrations: a seemingly good initial theory becomes downright poor once investigated. This is one of several reasons that at least one thing became clear to me after looking at the most important theories: the Voynich manuscript does not offer an obvious explanation. Nevertheless, it is possible to narrow down the solution. First of all, I am not aware of any convincing theory regarding the author of the manuscript. That the manuscript is a forgery made in the early twentieth century, which was a plausible and much-discussed hypothesis for decades, can meanwhile be ruled out. This assumption is not only disproved by the radiocarbon analysis but also by a recently discovered seventeenth-century document that mentions the Voynich manuscript. Therefore, we must search for the author half a millennium in the past. However, Roger Bacon and several other proposed authors (for instance Leonardo da Vinci) lived in the wrong time to be the author, and Antonio Averlino is a very speculative guess. This means that the author of the Voynich manuscript was probably an anonymous artist living in the first half of the fifteenth century. Maybe it was even someone who is completely forgotten today.
I am also not aware of a convincing theory explaining the meaning of the many figures in the Voynich manuscript. Neither the plants, which have no equivalent in nature, nor the other illustrations make any sense. The most likely explanation therefore is that the figures don't have a meaning at all. Probably they were just included to make the manuscript look more mysterious. If the figures have no meaning, it is very likely that the Voynich manuscript didn't serve a real purpose. My favorite explanation for the manuscript is that it was simply created to produce a mystery. Maybe the author intended to sell it for a large amount of money to a wealthy contemporary, or maybe he even acted by order of such a person. Another theory, which I consider plausible, posits that the Voynich author was a mentally ill person (for example, someone suffering from autism); it is quite common for mentally ill people to create art. As far as I know, this hypothetical origin of the Voynich manuscript has never been researched by an expert.
If the Voynich manuscript is really a hoax-as I suspect it is-it is very likely that the text is mere nonsense. However there are two other theories worth mentioning. The first one is based on a fact known by every cryptologist: that the design of a secure encryption procedure will become distinctly simpler if the encrypted text is longer than the original text. Under this condition, it is in fact possible to hide the original information in meaningless filler text. It is absolutely possible that the author of the Voynich manuscript used this trick. Perhaps he transferred an original, shorter text (e.g., 50,000 letters) into an unknown script and extended the result to the 170,000 letters he finally put down. Unfortunately nobody has yet discovered a pattern that allows the separation of the original from the filler letters. If such a pattern exists and is sufficiently complex, then there is only a minimal chance that it will ever be decrypted.
If the Voynich text is not simply nonsense, I consider the artificial-language hypothesis developed by William Friedman as a second plausible theory. It is certain that alchemists and scientists made efforts to develop secret languages during the Renaissance. Maybe such a secret language expressed in unknown letters forms the basis of the Voynich manuscript. However, there is no precise assumption as to what the underlying artificial language might have looked like.
However, the most probable theory is that the Voynich manuscript does not contain meaningful text. The papers of Gordon Rugg and Andreas Schinner suggest not only this theory of a hoax, but the assumption that the manuscript has no real purpose. In my view, there are still some interesting open research questions in this area. Was it possible to produce hundreds of pages of nonsensical text that has many things in common with natural language by using a method available in the fifteenth century? This question is heavily debated among Voynich scholars, and some think it impossible. However, there have been but few attempts to create a text that resembles the original. Rugg's method is one example, but there should be many more-involving encryption procedures as well as methods for producing meaningless letter sequences. If one of the results has statistical properties similar to those of the Voynich text, this might tell us which method the Voynich author applied.
The Voynich manuscript is and most likely will remain a riddle. We can hope that the manuscript will not merely become a playing field for mystics and pseudoscientists. After all, the subject is fascinating enough without adventurous speculation.
Note
The original German version of this article was published in the journal Skeptiker. This English version was translated by Susanne Kisser and edited by Skeptical Inquirer staff. In October 2010, the author updated this article with considerable new information.
References
Brumbaugh, Robert S. 1978. The Most Mysterious Manuscript: The Voynich “Roger Bacon” Cipher Manuscript. Carbondale: Southern Illinois University Press.
Feely, Joseph M. 1943. Roger Bacon's Cipher: The Right Key Found. Rochester, NY: n.p.
Kennedy, Gerry, and Rob Churchill. 2005. The Voynich Manuscript. London: Orion Books Limited.
Levitov, Leo. 1987. Solution of the Voynich Manuscript: A Liturgical Manual for the Endura Rite of the Cathari Heresy, the Cult of Isis. Laguna Hills, CA: Aegean Park Press.
Newbold, William Romaine. 1928. The Cipher of Roger Bacon. Philadelphia: University of Pennsylvania Press.
O'Neill, Hugh. 1944. Botanical remarks on the Voynich MS. Speculum 19: 126
Pelling, Nick. 2006. The Curse of the Voynich: The Secret History of the World's Most Mysterious Manuscript. Surrey, UK: Compelling Press.
Rugg, Gordon. 2004. An elegant hoax? A possible solution to the Voynich manuscript. Cryptologia 28(1) (January): 31.
Schinner, Andreas. 2007. The Voynich manuscript: Evidence of the hoax hypothesis. Cryptologia 31(2) (April): 95.
Klaus Schmeh, a computer scientist, works as an encryption expert for a German software producer. He is author of several cryptological and scientific books (mostly published only in German). His book Codeknacker gegen Codemacher (W3L-Verlag, 2007) deals with the history of encryption technology. He is a member of the German skeptic organization GWUP.
