I just uploaded the first part of the third episode of Voynich Talk!
You are not allowed to view links.
Register or
Login to view.
It is aimed at a broad audience, and attempts to tackle the issues with substitution ciphers in a way the layman can understand. I made this because I feel like there is a bit of a chasm between people who understand the statistics of Voynichese well (like many here on the forum) and those who don't. Bridging this gap is not an easy task, because concepts like conditional entropy can feel overwhelming to the mathematically impaired (it took me quite some time to understand it myself).
My hope is that next time someone exclaims "it's been solved", we can just have them watch a 12 minute video. This first part is all about the entropy problem, other issues may be tackled in subsequent parts.
I did put in a lot of effort in trying to up my video editing game, so I hope it is a bit entertaining for experienced researchers as well. In one scene I animated the crossbow, and there's a joke about corn.
Enjoy!
[
attachment=9171]
I think the most dangerous thing about the first contact with the VM is the sense of achievement.
A few words that sound familiar. A little research, a few more words, and it's done.
‘Can't you see it, it's clearly Polynesian’
Suddenly it doesn't work anymore and now it's forced.
The English were there, the French and the Dutch. That changed the script a lot.
The fact that Polynesia wasn't even known yet doesn't bother anyone.
Translated with DeepL.com (free version)
Aga: that's exactly right. This is also why I am convinced that the majority of theorists are intelligent people who were misled by the MS. Some of them write surprisingly learned papers. But upon analysis it always appears that they fell into the same trap. In the next video I will talk about this and matters of methodology. I will also solve a generated substitution cipher to explain the optimal approach and how it is opposed to all Voynich theorists' approach.
(10-09-2024, 11:20 PM)Aga Tentakulus Wrote: You are not allowed to view links. Register or Login to view.I would also like to thank Koen for the video. I didn't understand too much because of the English, but the meaning is clear.
I'll wait until the AI also does (spoken) language translations.
I hesitated to comment yesterday so as not to seem too demanding.
Indeed, the video is aimed at English speakers, because there are no subtitles, worthy of the name. Youtube no longer has a subtitle function?
The automatic subtitles talked a lot about cheese, I didn't understand anything, even by reducing the playback speed.
There is also no plan to be able to replay certain parts.
To end on a positive note, the length of the video is very reasonable, it is to be congratulated!
Great work, Koen! And the graphics are amazing...
I ran the video through You are not allowed to view links.
Register or
Login to view. (excellent online transcription software, with free trial) and fixed a few errors (Voynichese is not some kind of cheese). Now I need to cut them into small chunks and time stamp-them with Subtitle Edit to make a proper subtitle file for Youtube. I've done half already.
Koen G Wrote:Hello and welcome to this episode of Voynich Talk.
If you are watching this video, there's a good chance you've heard about the Voynich manuscript before.
You may also have heard that it's been solved.
And then solved again.
And again.
And again.
It has been solved many times by many different people in many different ways.
You may even have solved it yourself.
Today we will learn why all of these solutions are wrong.
Because despite all these theories being different, they also fall short in the exact same way.
That being said, there are a few theories this video does not cover.
For example, Timm and Schinner believe that the text of the manuscript is meaningless, and was generated through some predetermined process.
Similarly, Gordon Rugg argues that glyph tables and a grill can be used to quickly obtain meaningless text compatible with the format of Voynichese.
Approaches like these are not part of this video.
No, we are interested in the most common kind of solutions.
The ones you see in the media every couple of years, or enthusiastically proclaimed on forums, blogs, or even in peer-reviewed journals, that sometimes forget to peer-review.
Let's have a closer look at this average Voynich solution.
How does it actually function?
Most Voynich theories, way more than you think, are driven by two main mechanisms.
Substitution and creative writing.
Both are equally problematic.
I will explain exactly why substitution cannot work as a solution to the Voynich manuscript.
In a later video we will turn to the wonderful world of creative writing and broader issues of methodology.
Substitution, in its most simple form, is called: simple substitution.
This is when each letter in your original text, which we call the plain text, gets replaced or substituted by its corresponding letter in the ciphertext.
For example, you may have heard of a Caesar cipher, where each letter is shifted in the alphabet.
But you could also use a scrambled key.
The effect remains exactly the same.
Simple substitution ciphers are very, very easy to crack.
And I will demonstrate this in the next video.
But we can go much broader than that
and expand our definition of substitution beyond simple ciphers.
Converting your text to an alphabet the reader is not familiar with, in this case for example Greek, that is also substitution because each letter can be substituted by a different letter form.
But you could also try something more complex like removing vowels and only converting consonants.
Your code may become harder to crack, but it's still substitution, and still completely unsuitable as a solution to the Voynich manuscript.
Just to give you an idea of how pervasive this issue is, let's have a look at this list of Voynich solutions compiled by Tavi.
These are all people who — each and every one — are convinced they cracked the code.
Each in their own mutually exclusive way.
I started going through them in order to get an idea of the variety of solutions we're dealing with, but I gave up soon for two reasons.
One is that many theorists don't communicate clearly about their methods.
They act like they come down to us, mere mortals, with a mysterious black box full of secret knowledge.
Sometimes, to be honest, I get the impression that they hardly know what they're doing themselves.
The second reason is that when the method was finally revealed, it always turned out to be substitution.
Apart from the aforementioned exceptions, all Voynich solutions are based on substitution.
But the problem is that the Voynich text cannot be the result of a substitution cipher.
Perhaps the biggest issue is, as experienced Voynich researchers often point out, the text's low conditional character entropy.
This simply means that the characters tend to appear in the same order within a word.
That when you get one character, it's easy to predict the next one.
Let's try something.
I'm thinking of a word in English, and I'm telling you it starts with a P.
Can you tell me which letter is next?
It could be A as in park, E as in pet, I as in picture, L as in play, O as in poster, R as in print, U as in publish.
If we allow words that originally came from Greek, we can even include H, S and T as in photo, psychology and pterodactyl.
There are just too many options, several of them quite likely.
And it doesn't end there.
If P is the second letter in a word, then the letter following it is completely unpredictable.
We could move P to the third place, fourth place, any place, and it would still be able to be followed and preceded by a whole range of letters producing perfectly valid words.
Voynichese does not work like this.
If I give you this character, you know almost for certain that this is the next one.
Then, you can guess with surprising accuracy that one of these two tall characters is next.
Having arrived at this point, over two thirds of cases are followed by either one of these.
It goes on like that.
Additionally, many characters appear almost exclusively in certain positions like at the beginning or end of words.
Once you pick a character, there are not many options for the next, and the next, and the next.
It's all very rigid and predictable.
The implications of this are enormous.
As René Zandbergen points out in a paper published in 2019, he focuses on the writings of Gerard Cheshire, but the same problems exist for every possible substitution solution in any known human language.
Cheshire assigns this character to D.
Big mistake.
Now D can only be used in the beginning of words, and it can only be followed by this character, which he conveniently reads as O.
So in Latin, these common words can no longer be formed in Voynichese, because the character must be word initial, and it must be followed by the O-character.
Keep in mind that these are some of the most frequent words found in any Latin text.
The system cannot produce them.
I was about to grant that Cheshire's system could form the word domus, but it turns out that this is not the case either, which just shows how unforgiving the structure of Voynich text is.
Using the table we see that Domus should be written like this.
The problem is that when you have a series of these I-shaped minims in Voynichese, they are preceded by this shape.
So, while Cheshire's Proto-Romance can't live in a proper domus, they can at least build a domais.
I used Latin examples here, but the exact same problems would arise for any human language.
If your language has a D sound, you want to be able to use it in various positions, and at least you want to be able to pair it up with more than just O.
The problem is not the chosen language or dialect, the problem is Voynichese, and our insistence to keep treating it as a substitution cipher.
The fact that Voynichese letters have a preferred order within words is unlike anything we see in any known writing system.
It has more in common with something like Roman numerals, where MMXXIV is a valid order, but VXMIXM isn't.
A similar effect could be reached by alphabetically sorting the letters within each word, like this, but then more repetitive.
People will often object that maybe the statistical problems can be solved by special conditions, like using some obscure language or even a dialect.
But switching from standard French to any Romance dialect doesn't move the needle at all.
Both languages are equally distant from Voynichese.
It simply does not behave like human language.
Similarly, assuming an abjadic system, so a system without vowels, does not help.
Human language needs a certain amount of flexibility, which Voynichese is not able to offer.
What about abbreviations? Well, that's even worse.
Typically, abbreviation will leave out the repetitive and predictable parts, exactly because they are repetitive and predictable: that's why they can be abbreviated.
So if you abbreviate a text, whatever remains will be rich and varied, dense in information.
Voynichese is the exact opposite of that.
It looks like it's made up of parts that should be abbreviated.
Here's a concrete example of how bad it is.
Let's take this character.
In Herbal A it occurs a total of 1262 times.
Literally 99% of those occur at the end of the word.
And 95% are preceded by this character.
Of those endings 90% are in turn preceded by this character, resulting in this or this.
This means that assigning a unique phonetic value to this character is madness, since whatever letter you choose to correspond to it can now only occur at the end of words.
But, you might say, what if there is a language with certain sounds that only occur at the end of words?
Well, sure, but then those sounds would also always be preceded by fixed sounds, and those in turn preceded by yet another fixed sound.
Similarly, as soon as you assign a plaintext letter to this character, your chosen letter can now only occur at the start of words, and it can only be followed by whatever letter you picked for O.
The most Voynichese-like thing many of us are probably familiar with is the spelling convention that Q should be followed by U, as is the custom in several languages that use the Latin script.
For example, in English, we could easily assign the Q sound to Q alone, without causing any trouble.
The U is redundant and predictable, and it could easily be omitted.
It is merely a relic of earlier spelling conventions, but it does remind us a bit of the behavior of Voynich text.
And still, Voynichese is much, much worse.
For starters, QU is just one part of any language that uses it.
For Voynichese, predictable groupings like this form the majority of the text.
But even if we focus only on the use of Q and compare this to this Voynichese pair, the situation in Voynichese turns out to be much more rigid.
I pulled up some random text in various languages I had sitting around on my hard disk from a previous project.
Here's 10,000 words of Jacob van Maerlant's Middle Dutch.
He uses Q 16 times, and 16 times it's followed by U.
It's the most frequently found in forms of the adjective quaet, which means something like bad.
But elsewhere, it occurs in the third, fifth or sixth position in the word.
Turning to Pliny's Latin, we see that Q is expected to perform at any position in the word.
No matter which language you choose, whether it be Latin, Turkish, Chinese, Swahili, Esperanto, or the unique Proto-Judeo-Roman dialect spoken by two Ethiopian nuns on a desert island, the situation always remains the same.
As soon as you start mapping glyphs to sounds, you can no longer expect to be able to hear them express the language using the system of Voynichese.
You are locking yourself into a system that's much too rigid and predictable, unsuitable for a human language.
(11-09-2024, 07:53 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.I ran the video through Whisper-Zero and fixed a few errors (Voynichese is not some kind of cheese).
Thanks, Nablator, that was very useful for me.
Posts about Cvetka's theory have been moved to Cvetka's theory thread You are not allowed to view links.
Register or
Login to view.
[
attachment=9186]
Super nablator.
Now I can even have it read to me. Now it just has to be synchronised.
@Koen
What is the word at 8:34?
I hear "trench" but I don't understand it. "French" doesn't seem correct either... why French?
"But switching from standard trench/French to any Romance dialect doesn't move the needle at all."