(Please excuse the impersonality of what follows, but so many linguistic Voynich theories are popping up at the moment that responding to them all individually would be an even greater waste of my life than trawling through their sad attempts at ‘research’: sorry, hope you understand, etc.)

Dear linguistic Voynich theorist,

Thank you so much for your fascinating [1] and generally well-researched [2] paper. Unfortunately, it seems that in your enthusiasm to publish [3], you may well have skipped past some important details that would have presented your evidence, reasoning, and conclusions in a somewhat different light.

For example, your literature review somehow omitted to mention any of the fifty-plus [4] linguistic Voynich theories that had been published previously: only the most eagle-eyed of barristers would be able to highlight how these differed from yours to any significant degree.

I was interested to note [5] that you repeated the late Stephen Bax’s opinion (perhaps without even knowing that he was the source) that it is OK for linguistic Voynich theorists to disregard all previous statistical and analytical work carried out on the Voynich Manuscript’s text. However, given that almost all of that evidence and observation runs directly counter to your linguistic Voynich theory (and indeed Bax’s as well), it is hard not to draw the conclusion that you have been more than a little [6] selective. By stepping past all the practical difficulties with reconciling Voynichese with natural languages that have been pointed out from 1950s onwards by the Friedmans and many others, it seems as though you have taken a particularly blinkered view of the challenge involved.

As to what you think comprises evidence that supports your particular linguistic reading, I’m sorry to have to point out that neither optimistically plucking words from all manner of dictionaries nor running your fragmentary and non-grammatical [7] output through Google Translate for validation constitutes ‘evidence’ in any normal sense of the word. Instead, these merely show that you are willing to throw darts at a map bindfolded and then claim to have invented the satnav. [8]

Your attempted argument as to how Voynichese’s word-forms structurally map onto the plaintext forms you highlight would have been more persuasive had you looked for evidence beyond the two or three pages from the Voynich Manuscript you restricted your attention to. In reality, had you done so you would have realized that the ‘language’ apparently employed in the Voynich Manuscript varies significantly between sections, between bifolios, and also between different page and line positions (line-initial, word-initial, word-final, line-final, labels, etc): and it turns out that the tiny subset into which you put your time is not at all representative of the rest. So your supposed ‘translation’ fails to scale up in any way at all.

Finally: given that in your paper you were unable to sustain your ‘translation’ of the (supposed) plaintext language(s) of the Voynich Manuscript beyond a handful of somewhat optimistic [9] readings, and that this is almost exactly the same level of (un)convincingness that other near-identical linguistic Voynich theories manage, it is hard [10] to feel persuaded by your claims that you have “finally peeled back the veils of secrecy on this most mysterious of manuscripts“. Instead, it seems overwhelmingly likely that you have fallen headlong into the same shallow logical traps as pretty much every linguistic Voynich theorist ever.

At this point, it would be a wonderful thing to be able to say that despite some methodological flaws and over-enthusiastic leaps to conclusion, your paper has still managed to advance our knowledge of the Voynich Manuscript. But this would not be entirely true. [11] Instead, all you have actually achieved is wasting your own time along with that of everyone else unfortunate enough to read your miserable offering: ultimately, your paper is a bland and tepid mix of pseudohistory, pseudoscience and pseudolinguistics that moves us all backwards rather than forwards in any perceivable way.

Sorry, hope you don’t mind too much, best wishes, etc, Nick


[1] This is a lie.
[2] This is a bigger lie.
[3] i.e. “slapdash haste”.
[4] Perhaps even a hundred.
[5] This is an even bigger lie.
[6] OK, “obscenely”
[7] OK, “pathetically nonsensical”
[8] It’s a good job I toned this sentence down, the first draft was a bit too strong.
[9] OK, “laughable and utterly random”
[10] OK, “so close to impossible as to make no practical difference”
[11] In fact, this would be a lie big enough to blot out the sun.

17 thoughts on “An open letter to pretty much every linguistic Voynich theorist ever

  1. Emma May Smith on February 2, 2018 at 8:30 pm said:

    Most linguistic theories are no better than most cipher theories. The truth is that most theories which seek to explain the Voynich text are simply bad, whatever starting point they take.

  2. Emma: yes, I’d completely agree that all current Voynichese theories (whether linguistic, cipher, shorthand, or hybridized combinations) fail to satisfactorily explain the complex mesh of behaviours we see, and for the same reasons you’ve explored on your blog – LAAFU, and so forth – along with Neal keys, etc.

    But why is it that the headline-grabbing Voynich theories my correspondents keep culling from Famous Media Outlets are always nutty linguistic theories? It’s to their PR-savvy linguistic Voynich theorist originators that this page is primarily addressed, as I’m sure you had already worked out. 🙂

  3. Peter M on February 3, 2018 at 2:02 am said:

    Emma May Smith, I’ve read your article, so far is fine. But I think you forgot something essential.
    Basically, I have to say most people only look at a cryptogram in two dimensions.
    Thanks to Nick, I can link pictures directly, instead of writing 1000 words where hard to understand anyway.
    Let’s take a look at the font and disassemble it, and rearrange it.
    I emphasize that is only an example.
    Then it looks like this …


    I already have a number of individual letters. (The list is longer).
    Since I know exactly as an encryptor, the person sees only where he wants to see, I use a trick.
    Then it would look like this.


    Now I have twice as many letters in one fell swoop.
    Now I bring the combination back into play.
    Then like that ..


    He did everything necessary. He has his own sounds hidden. The system is easy and quick to remember. Final syllables not included. The man was definitely not stupid. He has kept all the rules. Final syllables not included.

    Nick, now maybe you understand me better, why I think nothing of the EVA.

  4. Koen Gheuens on February 3, 2018 at 5:00 am said:

    Isn’t substitution-abjad-anagramming a cipher solution? Or is any theory with a natural language output a linguistic one? I just don’t see many solutions of other types being proposed, apart from ones that are inherently unprovable. (Basically any “it means nothing” theory).

  5. James R. Pannozzi on February 3, 2018 at 6:50 am said:

    Dear Nick:

    Your most recent post on linguistic theorists has completely reassured me.

    I was beginning to worry for a while but now am certain that, thank goodness, you are as crazy as the rest of us.

    Please pardon me as I must return to my table of Mayan glyphs and continue to seek out the morphological transformations from them back to the Voynich script(s). I promise not to publish any papers until the ancient Mayan priest I’m channeling assures me that I have everything correct and that it’s a favorable baktun.

    Knowing you will understand, please be patient !


  6. Emma May Smith on February 3, 2018 at 9:34 am said:

    Koen: I personally see different kinds of theories according to how they explain surface patterns. A linguistic theory sees the surface patterns as a result of the underlying language. A cipher theory sees the surface patterns as a result of the encrypting method. So a simple substitution cipher would indeed be linguistic in this sense, but the overlap is otherwise small.

  7. Jackie Speel on February 3, 2018 at 11:35 am said:

    One of my suggestions is that someone ‘literate and fluent in writing in one script’ was trying to copy something in a script unfamiliar to them – so they ended up ‘writing what was effectively nonsense in a smooth hand, but based on something real.’ (There are many current scripts which subscribers could try the experiment with.)

  8. James: I wouldn’t dare interrupt anybody doing Good Science, so please continue as you were. 😉

  9. Davidsch on February 3, 2018 at 12:20 pm said:

    Consider that the contrary is also true:

    In the current climate, publishing *any* theory would need a huge amount of courage. As a result, a researcher that takes him/herself serious, will definitely*not* publish his theory, or any part of it.

  10. Davidsch: sadly, courage is not the same thing as wisdom or brains. 🙁

  11. Koen Gheuens on February 3, 2018 at 3:02 pm said:

    Emma: exactly, there is no difference between a substitution cipher and conversion to a different, existing script – apart from in the intention of the transcriber; does he want to obscure or rather open up the contents?

    But the most recent theory uses anagramming, which would make it a cipher solution in my view. Hence my question to Nick: when is something a linguistic solution and if you don’t like those, what would you rather see?

  12. Koen: what I’d like is for people to stop talking about cipher and language and shorthand as if they were mutually exclusive (they’re not), and instead to start putting the hard work in to figure out how all the different parts of the Voynichese ‘engine’ empirically work.

    Given that the vast bulk of the evidence suggests that what we are looking at is not a simple language or a simple cipher, I really wish people would stop trying to solve Voynichese in its entirety without bothering to do any of the foundational work first: people can’t just ‘wish’ their way to the chequered flag, they have plenty of driving to do en route. :-/

    Examples of foundational question that need answering:
    * How should Voynich be parsed?
    * How should it be transcribed?
    * Can we tell the difference between well-formed Voynichese and copying slips?
    * Can we reliably reconstruct what miscopied text should look like?
    * Is there a definite evolutionary sequence between the different Currier ‘languages’?
    * Can we discern the order in which the pages were composed?
    * Can we reliably transform the text of the different Currier pages into one another?

  13. Fred Brandes on February 3, 2018 at 4:15 pm said:

    I suggest that any budding Voynich theorist be required to read The Shakespearean Ciphers Examined before being allowed even a peek of the manuscript.

  14. Fred Brandes: an excellent suggestion! Even though it was published in 1958, it can borrowed from the Internet Archive here:

  15. farmerjohn on February 3, 2018 at 6:32 pm said:

    The situation with the VMS in many aspects is similar to one with Fermat’s Last Theorem IMHO (though problems themselves are incomparable; FLT is harder by a mile).
    Uspenskiy in his book (in Russian) tells story about one Russian (Soviet at that time) univercity (MGU I think; Moscow State Univercity) which was receiving so many “solutions” of FMT, that a special reply form was created: “The first error in your proof is situated on page ____ in line ____”. From another source I learned that this still wasn’t enough. There were wrong “solutions” which couldn’t be answered that way:)
    So while answering all VMS theorists at once in single sarcastic post may be a big fun, it cannot improve the current situation. If one really wants to avoid farce, each “solution” should be taken separately and the precise thing which is wrong should be pointed (it’s easy in most cases). This is very painful, very sad, but very true.

  16. D.N.O'Donovan on February 4, 2018 at 3:53 am said:

    Among the contemporary reviews of the Friedmans’ book, I like Virgil Whitaker’s (Stanford) – and he’s not responsible for mis-spelling “Elizebeth”

    Review of
    The Shakespearean Ciphers Examined. An Analysis of Cryptographic
    Systems Used as Evidence that Some Author other than William Shakespeare Wrote the Plays Commonly Attributed to Him.
    [authors] William F. Friedman and Elizabeth S. Friedman.

    Review by: Virgil K. Whitaker, Science, New Series, Vol. 127, No. 3304 (Apr. 25, 1958), p. 965. (JSTOR)


    It begins:

    ” The Friedmans have attempted a scientific examination of an irrational cult. That is at once the strength and the weakness of their book…”

    ending with (among other comments):

    “As a scientific examination of evidence, this book is devastating, and it would settle the controversy forever if a faith could be destroyed by rational argument.

    But the faith itself is, to the student of human behavior, more interesting than
    the spurious ciphers developed to buttress it. … A fuller discussion of the human phenomena behind the alleged cryptograms would have made the book, for one reader at least, more interesting and more significant as a contribution to learning.”

  17. Davidsch on February 6, 2018 at 1:20 pm said:

    from nick >>Given that the vast bulk of the evidence suggests that what we are looking at is not a simple language or a simple cipher, I really wish people would stop trying to solve Voynichese in its entirety without bothering to do any of the foundational work first: people can’t just ‘wish’ their way to the chequered flag, they have plenty of driving to do en route. :-/

    ! Kudos, Like and thumbs up !

Leave a Reply

Your email address will not be published. Required fields are marked *

Post navigation