The Voynich Ninja

This thread is triggered by You are not allowed to view links. Register or Login to view..

There are plenty of cribs in the Voynich MS. Here, I use the following meaning of the word crib:
"a plaintext word or phrase that is known or strongly suspected to appear in a cipher text."

If one wants to be careful, one could also call them potential cribs.

The point where such cribs appear does not necessarily have to be known, but the more interesting cases are those where we can have a strong suspicion what a particular word should mean.

In this first post I highlight two cases.

The first is from D'Imperio. Figure 30 on page 108 lists the 'houses of the moon' commonly also referred to as the 'mansions of the moon'. According to traditional astrology there are 28.
Now the astronomical / cosmological figure on You are not allowed to view links. Register or Login to view. shows 28 words each covering a segment of a circle.
This is as good a crib as one can get.
But it even gets better. D'Imperio's figure shows that the majority of these names start with an 'a'. In You are not allowed to view links. Register or Login to view. the majority starts with Eva-o or Eva-y. There are three or so that start with 's'  and there are also three or so in the MS that start with Eva-s.

The second one is one I proposed myself some two decades ago by now.
f67r2 shows a circle split into twelve segments. There is a good chance that these represent the twelve zodiac signs.
The second ring from the outside has short pieces of text covering 3-4 lines each.
In seven of 12 segments there is one word standing apart a bit.
There is a good chance that these are the seven planets ruling this zodiac sign.

There are plenty more cribs, which I may add later on. Others are welcome to do the same of course.

Quote:In cryptology, a crib is a piece of information which may lead to the decoding of a message.
-Patrick Lockerby, July 26th 2017-

Can anyone exemplify how such a crib is used to infer the plaintext ? Or is the idea that a uniform methodology is used wrong?

This definition of cribs from Patrick Lockerby is much more generic than the one I used.

Even in the stricter meaning of the word, there is no uniform method. It just becomes simpler when the suspecter cipher method is simpler.

Simple substitution ciphers may seem easy to resolve, but this is only true if the text is long enough, and one has a computer.
It is here that cribs are essential. They give away a few of the letters, usually high-frequency ones.
With homophonic ciphers, the gain from such a single crib is only a fraction of the gain in case of a simple substitution.

Modern encryption used for secure data transmission is fully crib-resistant. This is because a would-be interceptor has the capability to generate any number of plaintexts, and then see how these are encrypted. Still, with all that information he should not be able to decrypt any other encrypted text.

I can imaging kind of a continuum of how handy a crib is, depending on a number of factors:

Scenario 1, maximum certainty

You know the underlying language (let's say Latin) and you suspect simple substitution. In this case, even if your crib is a single word, it will be incredibly useful. Say you have a picture of the sign Aries, and with it ciphertext Bsjft. Congratulations, your cipher is now solved.

Scenario 2, uncertain language

This is the same as scenario 1, only now you don't know the underlying language. You can start testing your crib by plugging in "Aries", but also any vernacular versions that seem plausible (Widder, Bélier, Ram...) and spelling variations. As soon as you find something that works, you will also need to test this system with other cribs. As long as you are dealing with simple substitution though, a few words should be enough to get you started.

Scenario 3, uncertain crib

You suspect simple substitution but your crib is a sentence and you don't know the exact phrasing. I think this might be more difficult than scenario 2, because there can be variation in terms of the words selected, syntax, word order and spelling. However, if you do know the underlying language, you should have a decent chance of success even without using computers.

Scenario 4, uncertain cipher

Imagine you know the underlying language and are certain that you are looking for the word "Aries", but the ciphertext is "aarbicedse". In this case, you could still either solve it manually or use a computer to brute-force a solution. Some ciphers will probably be unsolvable this way, though I don't know enough about it to know whether this would be the case for any of the standard medieval ciphers one would encounter.

Scenario 5, combination of uncertainties

All of the above, apart from scenario 1, come with their own set of difficulties. But now imagine you are not certain about the exact wording of your crib, you don't know the underlying language and you don't have a clue about the type of cipher you're dealing with. You don't even know if the meaning of your ciphertext is retrievable (one-way cipher) or if it contained any meaning to begin with. You also don't know if your text describes the drawings in a straighforward way, or in which way it relates to them, if at all. That's the situation we're in.

This is an excellent and very interesting subject. I thought it worth listing subcategories of crib:

1) A given Voynichese word corresponds to a given meaning. So we can say with confidence that this Voynichese word has precisely this meaning, though we may not know the underlying language in which the word, enciphered or not, was written. This is the gold standard of crib.
2) A given set of Voynichese words correspond with a given set of known meanings though the precise mapping between the sets is not known.
3) Somewhere in a piece of Voynichese text we expect to find a word with a given known precise meaning. The usefulness of this depends on the length of the text.

(This presupposes the idea that Voynichese spaces correspond to real spaces and therefore Voynichese words correspond to real words)

And of course of fundamental importance with all this is the degree of confidence that we can have in the crib association. Roughly what chance do we think there is that the crib is reliable.

So when listing suggested cribs I think it worth stating what type they are and very roughly what level of confidence one has in them.

I seems with the mansions of the moon example the claim is presumably that a cyclically ordered set of Voynichese words corresponds to a cyclically ordered set of meanings, though the precise correspondence is not known.

I don't know anything about the mansions of the moon, but I found this image.

And this is the corresponding Voynich manuscript mansions of the moon claim.

I wonder if the mansions of the moon were named differently in different languages. In the inner namings they almost all begin "Al" i.e. "The" and so appear to be Arabic.

Are the positions of the 1st and 28th item arbitrary or is it possible to come up with a definite 1 to 1 mapping?

Each mansion appears to have a name and an indicator star. So one would have to check which Voynichese word corresponds to which name or which indicator star.

It is not clear to me whether the mansions of the moon are always named the same or have the same meaning given different languages and cultures.

According to the following there are quite a few possible text interpretations, which makes creating a one-to-one mapping more difficult:

You are not allowed to view links. Register or Login to view.

I think what would be useful is to find a manuscript illustration which more closely corresponds to ours; a lot of the illustrations online appear to be from Arabic texts.

I find this idea for a crib potentially very interesting, in particular as there are 28 entries and therefore it could constitute a large crib, however clearly there remain a few issues to resolve before coming up with a mapping of Voynichese words to meaning that one can be confident with. Also I don't know enough about the subject to say with confidence that You are not allowed to view links. Register or Login to view. does indeed represent the mansions of the moon.

As for the uncertain cipher, I would assume that for the relevant time frame a not so simple substitution ( see following description ) is a possibility. Assuming Latin, one could take repeating letter pairs (A-B-A) as crib, search for them in a comparative text ( here regimen sanitatis ) and check if such pairs occur in whole words in the VMS ( e.g. "SHEDY QOKEDY SHEDY" <> "ti-ga-ti" ). From this, one could derive a word or series of words in the best case.

Even if my attempt from 2018 was not successful, I think the method is promising.

see: You are not allowed to view links. Register or Login to view.

edit: Of course, I am not 100% sure if such a " decomposition " of words was even conceivable at the given time frame.