VOYNICH PAPER

This is a bit of an unorthodox post, but I recently wrote a paper critiquing Stephen Bax's analysis of the language in the Voynich manuscript, and I just wanted to post it here on my site for posterity.

One of the more recent theories concerning the Voynich manuscript was proposed by Stephen Bax, who was a professor of Applied Linguistics at the University of Bedfordshire until his death in 2017. In 2014, he announced that he had “decoded 10 words and identified the approximate sound values for 14 symbols” using the same “bottom-up” approach that was used in deciphering ancient Egyptian and Linear B (“Decrypting the Most Mysterious Book in the World”). This theory, developed in his free time, generated him considerable media attention and still is one of the more accepted arguments over the codex, despite some blowback from the linguistic community.

Using the assumption that the book was written in a natural language, Bax concluded that the first words on every page with a plant illustrations stood for the name of the plant. After a multitude of dead ends, Bax identified one word that he was fairly certain meant “juniper” because it was next to what he said was a juniper plant. By analyzing terms for “juniper” in other languages (including Hebrew, Arabic, and Georgian) and noting letter alteration, he found the word to be /arar/ (Bax 17).

Now Bax had two letters to work with. In one of the star charts, he found a constellation cluster that he was fairly certain was the zodiac Taurus. This had the /a/ and /r/ letters from before, so by comparing the Voynichese word with the Arabic name for “Taurus” he identified /taərn/ as the pronunciation (Bax 20).

Next, Bax found an illustration on folio 41v that he claimed closely matches the leaves of the coriander plant. The first word on that page had the /a/, /ə/, /r/, and included three other letters, two of which he assumed to be /k/ and /oo/, and one of which he didn’t identify. After providing a “wide range of names” for coriander in other languages to show that the word varies a lot, he identified the perceived label as /kooratu?/ with the question mark being the unknown letter and the schwa becoming a /u/ because the letter for that could apparently stand for both sounds (Bax 25).

Working with  /k/, /n/, /t/, and a new assumption for an /ir/ letter, Bax extrapolated /kntuirn/ as the name for the Centaurea genus of flowers, and then with the same method found /kntuirn xəur/ to mean “the Centaur Chiron”  (Bax 29). The word /xəur/ had yet another symbol for the letter r, which Bax acknowledged as “anomalous” but explained it as probably being an allophone of the phoneme.

For the remainder of the paper, Bax continued to use the same method of building off his previous discoveries and identifications of plant types to pinpoint a word pronounced as /kaəur/ and meaning “black hellebore” by using the Google search engine to compare it to an Indian word for hellebore (Bax 37), then find /kaəur xar/ as meaning “black cumin” (Bax 43), /kooton/ as meaning “cotton” on a folio where he’s “not certain whether this page depicts cotton or not” (Bax 47) and /ksar/ as meaning “Indian crocus” (Bax 50). 

To conclude, Bax iterated that Voynichese must be a natural language and not a hoax nor a cipher, and that the language is not Indo-European. Rather, there is “possible Near/Middle Eastern influence” in addition to a “European element”, “some influence from the Indian subcontinent”, a “Caucausian influence”, and a possible “Turkish influence cannot be ruled out” (Bax 52). He does not establish anything concrete about the author’s identity, and he says his work “suggests that the Voynich script could have been created by a small group wishing to encode a previously unwritten language or dialect.” Finally, Bax acknowledged that there were a lot of flaws in his research but he hoped that it would be a useful contribution to Voynich scholarship (Bax 53-54).

Obviously, there are numerous flaws with Bax’s work. One of the major issues is that the geographical comparisons are incredibly inconsistent. Bax mentions similarities to four different language families, including languages like Italian, Arabic, Hebrew, Georgian, Turkish, and even English. It is highly unlikely that Voynichese could have drawn on all of those sources; it is far more probable that many of Bax’s identifications are erroneous, and that he was making unreasonable conjectures through shoddy linguistic analysis. Furthermore, the paper that he relies on for his plant identifications claims that the botanical evidence matches an Italian origin (Sherwood 1), and the illustrations generally match a medieval European style rather than something from the Middle East, so there’s further contradiction there.

This potential issue exacerbates another major problem with Bax’s method. Many word identifications depend on previous words being correct: for example, /taərn/ needs /arar/ to be right for it to be right, and /kooratu?/ needs /taərn/ to be right for it to be right. Bax readily admitted that “each of the above readings of /arar/ and /taərn/ is insubstantial and must be considered speculative”, which effectively means that the whole paper is speculative and not grounded in concrete linguistic analysis.

This is especially bad because the problems with Bax’s assumption about /arar/ are manifold. It seems like he draws the word out of thin air, building the entire framework of his argument on it. The connections to Georgian, Arabic and Hebrew are weak, based solely off the need to find a term with an ABAB structure. The word also occurs several more times in the manuscript - too many to reasonably be interpreted as “juniper” - and Bax noted a prefixed p and t to the term, which, in his own words, “renders the interpretation uncertain”. If that’s not clear, what is?

Other words seem even less likely to be accurate. For instance, the classification of /kooratu?/ as meaning “coriander” was particularly suspect due to the unidentified letter, the amount of time Bax spent listing out words from unrelated languages to prove his point, and how most other linguists agree that if there is a natural language involved at all, “the words of the Voynich Manuscript should be split into two character syllables” (Bloem 15). 

Moreover, there is further inconsistency in the identification of three different letters for /r/. Bax explained this by saying that one of the letters is for an /ra/ sound, one is for just /r/, and one is for either /ur/ or /ro/ (Bax 30). He struggled to justify this, proposing the possibilities of variations based on position, case, or allophony, but he clearly wasn’t able why his theory “flies in the face of supposed common sense” (“Stephen Bax and the Voynich Manuscript…”).

Another underlying problem with Bax’s work is the tentativeness of his plant identifications. Contemporary thought on the artwork suggests that many of the plants are fake or highly altered through centuries of being copied down from different manuscripts, but Bax said he knew what some of them were “with some confidence” (Bax 12). Bax relied heavily on a paper by Edith Sherwood, who admitted that “some of [her] correlations may be incorrect with respect to either genus or species” and “in a few cases the drawing has insufficient data to identify the plant” (Sherwood 1). Sherwood has wildly vacillated between several theories in the past and been criticized by multiple academics for her tenuous conjectures, but Bax treated her paper as the gospel upon which he could base his linguistic analysis.

Furthermore, the idea that all the first words on pages with plant drawings are labels has an “immediate problem” despite looking reasonable on the surface: “almost all of the Voynich’s Herbal A pages” begin with one of four characters, /arar/ being a rare exception (Stephen Bax and the Voynich Manuscript…). That seems extremely sketchy if Bax is to be believed.

Finally, the work is affected by Bax’s personal bias. This was largely revealed through an AMA (“Ask Me Anything”) post that he made on Reddit, where he informally revealed his opinions on the Voynich Manuscript. “Don’t tell anyone OK?” he wrote in one comment. “I think it is probably an invented script, probably by a small group trying to study and pass on knowledge, maybe in a region not far from Europe, e.g. Turkey Iran, Caucasus. I then think that for some reason - war? - the group died out. But not much proof ….. YET.” Elsewhere, he essentially admitted that he wasn’t sure whether his ideas were right and he only posted them online to get peoples’ reactions to them. Between these and other such comments, Bax’s AMA shows that he had preconceived notions about the Manuscript which he was trying to retroactively prove. He even declared “I’m biased!” in another comment, which doesn’t exactly reinforce his paper with scientific merit (“I am Stephen Bax, researching the Voynich manuscript”). 

​In conclusion, Stephen Bax’s research has a litany of issues, including geographical inconsistencies, dependency on past translations, three versions of the letter r and inconsistency in letter identifications, the assumption of plant pages having initial labels, tentative plant identifications, and personal bias. While he may be all or mostly wrong, he still brought an important new perspective to the study of the Voynich manuscript which may help future analysis. It’s a lot harder to propose a new theory than to debunk an existing one, and Bax made an overall useful contribution.

Sources

Bax, Stephen. A Proposed Partial Decoding of the Voynich Script. University of Bedfordshire, Jan. 2014, http://stephenbax.net/wp-content/uploads/2014/01/Voynich-a-provisional-partial-decoding-BAX.pdf. 

Bax, Stephen. “I Am Stephen Bax, Researching the Voynich Manuscript.” Reddit AMA, 24 Feb. 2014, https://www.reddit.com/r/IAmA/comments/1yt8cb/i_am_stephen_bax_researching_the_voynich.

Bloem, Peter. Unsupervised Analysis of the Voynich Manuscript. University of Amsterdam, 3 July 2006, https://staff.fnwi.uva.nl/b.bredeweg/pdf/BSc/20052006/Bloem.pdf.

McCormick, Richard. “Decrypting the Most Mysterious Book in the World.” The Verge, 28 Feb. 2014, https://www.theverge.com/2014/2/28/5453596/voynich-manuscript-decrypting-the-most-mysterious-book-in-the-world. 

Pelling, Nick. “Stephen Bax and the Voynich Manuscript...” Cipher Mysteries, 21 Feb. 2014, http://ciphermysteries.com/2014/02/21/stephen-bax-voynich-manuscript.

Sherwood, Edith, and Erica Sherwood. “The Voynich Botanical Plants.” The Voynich Botanical Plants, 2008, http://www.edithsherwood.com/voynich_botanical_plants/index.php. 
​
Voynich Manuscript. Yale University Beinecke Rare Book and Manuscript Library, ca. 1401 - 1599.