Logic is in the Eye of the Beholder

“Komm mir nicht mit Logik: Mit Logik kannst du alles beweisen.” — “Ja, aber nicht jeder.” (Me and a colleague of mine)

Lately, the marginalia of the VM have received more attention again, last but not least due to Dana Scott’s project to give them a thorough examination.*) You remember, the marginalia are the apparently latin letters on several of the VM folios, which, for the better part, are readable, but refuse to yield sense whatsoever. Now, Ye Olde Nick Pelling has presented his own theories to explain the features of the marginalia, which runs as follows, IIUC:

  1. Once upon a time, the author of the VM wrote down the marginalia in plain text. We may only speculate what he wrote, but, as marginalia are wont to be, it was probably the least cryptic part of the VM.
  2. Over time, these writings got weathered, and were only poorly legible anymore.
  3. Some renaissance or baroque whippersnapper came along and decided to emend the marginalia, “restoring” their meaning. Unfortunately, he had no idea what this meaning was, and erred considerably.

This is why we today are left with unreadable strings of gibberish alongside the regular VM text (which are unreadable strings of hopefully meangingful content).

Now, while not wanting to defame Nick’s merits regarding VM research in general I’m under the vague impression that here he has fallen victim to a logical fallacy. One question is, why should the marginalia have degraded so much more strongly than the regular text, which is still in reasonably good condition. (Even the VM letters within the marginalia of f116v are still quite legible.) But this is forgivable. The much more serious question is:

Why don’t the emendations make sense?

If we assume Mr. Whippersnapper restored what he thought was the original text, why is it anchiton oladabas, of all possibilities, rather than “Kilroy was here”, or “The King is a fink”? No matter what the state of the VM was when Monsieur or perhaps Signore Whippersnapper (vielleicht Herr Jungspund) decided to intervene, the current state of the marginalia is good enough for the better part of the letters to be clearly legible, ie if there was an emendation, we should still be able to read and understand it.

I’m afraid, one of the reasons Nick might want to stick to this theory is — knowingly or unknowingly — because it introduces a “meat grinder” in the interpretation, a technique all too common in VM studies. The trick then is to use a process of any kind which will convert some input to output, while introducing “degrees of freedom” in the interpretation.

Degree of freedom” in this context means the chance to arrive at different outputs starting from the same input. A notorious example would be the use of anagramming, which allows one to arrive at a lot of different possible solutions from the same input (in this case the ciphertext), with no means to distinguish what the correct of these solutions would be. The more degrees of freedom any interpretation or decipherment introduces, the more arbitrary the results become, and the more possible scenarios exist with which the results would fit in.**)

In this case, if we take the marginalia at face value, we’ll be hard pressed to make sense of them. But if we say, “Hey, there’s been some correction, and here a wrong addition, and there …”, voilá, we arrive at legible text. Unfortunately, at pretty much any legible text that suits our fancies…

*) Not to be so easily bested, I decided to launch a similar project myself which I had pondered for some time, but never mustered the energy to do…

**) I feel I have overextended my grasp of English grammar with that sentence. Bear with me, please.

Advertisements

13 thoughts on “Logic is in the Eye of the Beholder

  1. Hi Elmar,

    Actually, there is no logical fallacy in my argument (well, in this part of it, anyway). For the marginalia, I closely observed at first hand the multiple inks, multiple handwritings, and emendations – and then worked back from what I saw to suggest a plausible reason why I thought the emendations came about in the first place.

    So, the logical fallacy here is actually that you think the presence of the emendations is the explanation, when I think it is an observation.

    My suggested explanation for the reason for the presence of the faulty emendations is that the marginalia were originally written in Occitan, a language I believe the later owner wrongly believed that he/she could read, and tried to correct.

    “Meat grinders” in cryptography I distrust just as much as you: but the default position on these marginalia should be that were not cryptographic. Rather, they are caught up in a tangled linguistic web, one that a little clear thinking (and a lot of good quality multispectral scans) should be able to de-layer and sort out. :-)

    Cheers, ….Nick Pelling….

    PS: passwords, schmasswords! ;-)

  2. Actually, I had planned to present the article to you before publishing to see whether I had misrepresented your theory (hence the “protection”), but I reckon this point is moot now. Well, at least that’s one cryptographic challenge you mastered. ;-)

    As for the logic of it all, you’re not seriously suggesting somebody mistakenly assumed he was capable of speaking a language? What frame of mind would that suggest? “I’m not sure if I speak Occitan… I’ll just give it a shot…”? (We’re not speaking of the occassional glitch here…)

    Whatever the “corrector” thought, he would have made emendations into something he understood. At least vaguely. And if that was German, Occitan, or Swabia-Silesian, we should be able to understand some of it at last.

    (I’ll grant you the point that I don’t have a sensible explanation for the whole marginalia business either.)

  3. Although I’m not convinced of the actuality of amendments to this marginalia, it if was, there could be other reasons it was done. For one thing, why assume the person was trying to make it understandable? I think it far more likely that if it were altered, it was to make it less understandable. IMO, there is a possible precedent for this, in that “octobre” and “libra”, and other zodiac names in the VMs, seem to be obliterated… as though someone wanted to erase some of the better clues. I would concur that seeing what is under any alterations would be valuable, but would not be surprised to find, rather than Voynichese, something more understandable to us… Rich.

  4. Hi Elmar & Rich,

    The emendations are there for all to see: you honestly don’t have to be much of a palaeographer or codicologist to pick up on the different letter shapes, writing strengths, ink colours, etc at play there. And a multispectral scan would be likely to reveal yet more subtleties to explain.

    Don’t forget that Occitan, which I suspect was the language originally used for the marginalia, was already fading fast in the 15th century. Yes, I too don’t feel particularly comfortable with the idea that it was later owners’ attempts at corrections that is making this hard to understand – but something happened to make them the way they are: and, like it or not, that’s probably the best explanation for these phenomena we currently have.

    And don’t also forget that there appears to be nicely-written Voynichese threaded in with these marginalia, which surely makes it more (rather than less) probable that it was indeed the original author who added these. Only a few words, but valuable ones!

    Really: for all their unreadability, the marginalia d give us one of the few pragmatic ways to explore the early history of the VMs. And so people who actively choose to ignore them or sideline them do their other research a grave disservice.

    Cheers, ….Nick Pelling….

  5. Nick, I fully agree with you that the marginalia could prove as the key to breaking the VM cipher. (Well, not “key” as in “cryptographic key”… more like a crowbar.)

    Which makes it just the more important to understand what’s really going on there…

  6. For what it’s worth…
    Nick proposes in his book on page 30 that, at the
    end of the marginalia on f17r, there is some
    completely faded text in Voynichese script.

    I can confirm that this is there and it is real,
    having seen it under UV light.
    Apart from that, I have no strong opinion on the
    level of emendations to the marginalia, let alone
    the time frame(s) in which they would have been
    made.

    Cheers, Rene

  7. Hi Rene,

    It’s not entirely true that the f17r Voynichese marginalia are “completely faded”: you can actually just about make out one or two of them from the Beinecke scans (particularly an “o” character), which explains why I was looking there at all. A Wikipedia editor would probably describe it as “close to completely faded text”. :-)

    We do all really need to get to grips with the emendations (if that is indeed what they are) – it’s a shame that the Beinecke turned down my specific multispectral marginalia scan proposal back in 2006, but perhaps things will be different in 2009 / 2010… :-)

    Cheers, ….Nick Pelling….

  8. Has noone considered that the marginalia are also in cipher?

    Sddms thd mnst lngkcbl fxplbnbtlnn tp mf.

    K.

  9. Frankly, I haven’t heard of anything like that, but of course you are right; this would explain why they can’t be read.

    The closest to it is the assumption that the marginalia (those on f116v at least) were an aborted decryption attempt (which has been mentioned several times). Of course, this would fail to explain why the marginalia seem to tie in with the illustrations, which in turn appear to be integral parts of the VM itself.

    What could be the motivation to use different encipherment schemes for the text body and the marginalia?

    (OTOH, it would be nice to explore whether the f116v marginalia could be explained as simple substitution ciphers from the VM ciphertext. If that was the case, the “abortive attempt” would suddenly become much more probable again.)

  10. Alas, the point about ciphers is that they need to be clearly written in order to be reliably deciphered – even though we can’t decipher it, Voynichese is actually a fine example of a clearly written ciphertext. :-)

    Yet the f116v marginalia can’t even be transcribed – it would take you days to list all the variations of “michiton oladabas” that have been proposed over the years.

    Even a fairly superficial examination of f116v should convince you that a number of different inks, nibs, and writing styles are present – certainly three (and arguably four) contributions.

    The presence of Voynichese on both f116v and f17r would certainly seem to indicate that cipher is the *subject* of the text: but codicologically, it’s perhaps a bit early to be wondering whether it’s enciphered as well as unreadable. :-)

    Cheers, ….Nick Pelling….

  11. Hi Elmar: I came across this post again today. You are familiar with the following points, as they have been discussed at length on other forums. But I wanted to put them here, to update my own ideas on the marginalia:

    1) Marginalia is usually understandable, at least to some degree, when found in other works. Even if all words cannot be read, the general intent, or the language it is written in, or something about it, is discernible. So then in the VMs case, considering we have at least three different sets of marginalia, from assumed different times, by different people, in styles somewhat different than the (also unreadable) main text, I have to ask “what are the odds?” that all of it is also unreadable? I think this implies that it was never meant to be readable in the first place, and is only there for “effect”, for decoration, intended to give a sense of authenticity to a fake work.

    2) The McCrone report determined that the ink of the main text and the marginalia on the last page are the same. To put this into context, they also determined that the ink of the page and quire numbers they tested were different from each other, and different from the main text/marginalia ink. But this then means that the marginalia was penned at the same time… probably by the same person, since it is probably from the same batch… as the main text scribe. But since the marginalia seems to be of a different style, with many different characters than the main text, it reinforces #1, above: That it was penned only for effect, to look like marginalia, to add a look of authenticity to a fake work.

    3) “Pox leber”: Many have noted that the last page marginalia seems to have the words “pox leber” in it. Famously, Johannes Albus spoke at the 100 Conference in Frascati in 2012, pointing out the possibility that the last page marginalia is a recipe, with goat liver, partly because of this. However, it turns out that 19th century scholars, in finding “pox leber” in a mid-16th century Hans Sachs poem, were perplexed where Sachs may have gotten the term from… and they determined that it did NOT predate the poem. And today, with the vast resources of the internet, we cannot find earlier versions, either.

    So given all the above, we have clear problems: To be authentic, and from the early 15th century, the Voynich author must have penned marginalia in different writing than he/she used for the text, and used a phrase from 150 years later. Or, if one wants to say it is not “pox leber”, they must explain, still, why the author would pen such marginalia in a different style. Or, if one will not give up “pox leber”… and some still hold that is what is probably written… they must claim it probably existed earlier, we just cannot find it.. running counter to the usual standards of research, and so on… then, still, needing to explain why the author wrote this odd stuff in the same ink. Or, say the McCrone ink results are wrong, or being misinterpreted by us, or that by chance a later marginalia author happened to mix ink identical to the main text ink.

    And many different versions of the above, using various elements, some contrary to each other, to explain these problems, and support the belief that the Voynich, and the marginalia, are genuine. But I think this is a very powerful smoking gun: Simply, these problems can all be explained, quite simply: The Voynich is fake, and the marginalia is, also, fake, and both were written post 16th century.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s