Could the Voynich Author Be a "Ramanujan-style" Anomaly?

[ololololo]

a You are not allowed to view links. Register or Login to view., and the Naibbe cipher

It sounds something like “I don’t know of anything more suitable for a VMS text than nonsense.”

Ignorance doesn’t mean it doesn’t exist. Perhaps the author was just a bit more perceptive than you.

[JoJo_Jost]

I know of only two encryption schemes that could do that: a You are not allowed to view links. Register or Login to view., and the Naibbe cipher. 

In know 3

[Bluetoes101]

Hi Chenxiang, 

For only 16 you seem very mature and intelligent for your age.
I think you are 100% correct, the VMS is truly unique so the person(s) behind it are likely unique also. So your intuition is correct (I think). 

There are lots of counterarguments to the broader theory, as you acknowledge. 

Some of the fun of the Voynich Manuscript is that it is a "sandbox" for ideas and tests, another is that researching and learning about "why you are probably wrong" can lead to a lot of learning about various things of interest and you will find you learn more than you would by being "the stubborn rock in the sea defying the waves" and refusing to look past your idea. These things you learn usually develop your ideas further, or snuff them out and lead to other thoughts and ideas.. the cycle continues, and the more you learn.

Specifically on the idea that the VMS was the work of one person I would recommend Lisa Fagin Davis "Many Scribes Theory". Secondly I would recommend Koen's work on A/B plants. I can't seem to find the links to them now, but hopefully you can find / someone can put it here. Anyway, the evidence put forward seems to show that we have something like 6 scribes and at least 2 "artists", maybe this number is 6 also. The point is that several people surely worked on the Voynich manuscript we see today, it doesn't mean there can't be one author, or that it is not a copy. However it does mean "one persons work" is a theory, and multiple peoples work on a manuscript is what we are looking at today. 

I hope that was some food for thought at least and a nudge into the adventure that is Voynich research. Best of luck!

[Jorge_Stolfi]

a You are not allowed to view links. Register or Login to view., and the Naibbe cipher

It sounds something like “I don’t know of anything more suitable for a VMS text than nonsense.”

The SPS=SBJ claim aside, I still see many arguments indicating that the VMS is NOT nonsense.  But I have also see many reasons to believe that it is NOT encrypted either.  Including those I posted above.

Ignorance doesn’t mean that [an encryption with the required proeprties] doesn’t exist. Perhaps the author was just a bit more perceptive than you.

The encryption method must generate a text whose "words" have the known "multi-slot" structure, the observed compact length distribution, and the observed Zipf-like frequency distribution, with some words specific to each section or page while others are more or less equally common in all sections.  

These requirements mostly rule out alphabet substitution ciphers, including those with one-to-many letter maps and nulls.  

A codebook cipher is possible because it would preserve the word frequencies and correlations of the original text, and generate compact "words", but would eliminating any glyph-to-letter correlations (thus making the encryption resistant to the methods used to crack substitution ciphers).  However, the word codes should have the multi-slot structure we observe.  Which excludes a positional number system, but allows a system similar to Roman numerals.

Another alternative is an encryption which does not preserve the original word divisions; that is, which hides the original inter-word spaces and inserts new ones.  The Naibbe cipher is an extreme example where each letter is encoded as a separate word.  It could also be a similar scheme using syllables instead of letters.  But, besides the practical problems that I mentioned in the previous post, the letter or syllable encoding method would have to yield "words" with the properties that we see in the VMS, including the vocabulary size, the compact word length distribution, the slot structure, and the Zipf-like "word" frequencies.  These properties seem unlikely to be satisfied by lucky accident; they seem to require deliberate non-trivial rules in the encryption method.  But 1450 was still a couple of centuries before the invention of statistics and the discovery of Zipf's law...

All the best, --stolfi

---

I know of only two encryption schemes that could do that: a You are not allowed to view links. Register or Login to view., and the Naibbe cipher. 

In know 3

OK...   
All the best, --stolfi

[Jorge_Stolfi]

For one thing, because such a cipher would have been cracked long ago.

Imagine that I wanted to complicate the homophonic cipher with nulls by doing the following:
1). By assigning various numerical values to the letters
2). Marking vowels in words as two symbols separated by a space.
3). By distributing the letters within the word into anagrams.
4). And having created a second alphabet, he applied it to his own cipher wheel.

The more complicated the cipher, the harder it is to decode. The crypto camp has never been able to give a plausible justification for why the Author would have felt it necessary to encrypt a whole book like the VMS with a military-grade cipher.  If it is indeed a herbal, why make it so hard to read?  If it is some dangerous "heretic" text, why waste all that vellum with those supposedly "decoy" images that would only attract the attention of the "inquisitors"?

And, anyway, it is unlikely that a complex cipher would produce a text with all those properties that we see in the VMS, that are characteristic of a natural language -- vocabulary size, Zipf-like word frequencies, ~10 bits of entropy per word, a common core vocabulary plus section-specific words, single-word labels, etc.  

For instance, an encryption that used  null letters or a position-dependent letter mapping like Vigenère would result in one word type W becoming several distinct word types W[1], W[2], W[3]...; and that would usually destroy the Zipf property.  To preserve it, the encryption method should be such that each alternative W[k] is chosen with a different probability that is just right.

Also, this argument “they would have cracked it long ago” doesn’t work. For example, there’s Cryptos — a somewhat more complex version of the Vigenère cipher. The last text on the tablet hasn’t been deciphered yet (even with the clues).

A Vigenère cipher and other position-dependent letter substitution schemes are much harder to crack than position-independent letter substitution or scrambling scheme.  But they have the above problem.

There are Bale’s cryptograms. One of the texts was cracked manually (!), and in our world of technology, it would be a piece of cake to crack it. Maybe it’s not a cipher after all?

The Bale ciphers are a codebook cipher, using some other text as the code book.  To crack the other two letters one must guess which text was used as the code book. 

But they stink of a con job from miles away...

All the best, --stolfi

[ololololo]

Jorge, I think all your statements are correct now, and I have nothing to object to. But keep in mind — the Naibbe cipher shows that, theoretically, it is possible to create a similar VMS cipher, which means that everything is not so straightforward...

[Jorge_Stolfi]

And yet, the drawings in the astronomical section have a sublime original artistic feel to them. It is so incredibly strange.

But it looks like the Artist bungled some of them:

• on f68v3, the wavy line was probably meant to be a circular cloudband (as in that Oresme "Genesis" diagram).  But it became an irregular octagon that looks like the edge of a thick plate or of a pool.
  
• on f69r, the number of "cups with stems" was probably meant to be a multiple of 3, and every three "cups" apparently were meant to be associated with one of the 16 sectors of the outer text ring.  At the same time the spaces between the stems were supposed to alternate between blank and filled with a text line.   To achieve both purposes the number of "cups with stems" should have been 48. But, as in several other diagrams, it seems that the Artist did not even try to divide the space evenly beforehand.  Maybe he did not even think that he had to draw a specific number of "cups".  It seems that he just drew one "cup" after the other until he closed the circle after 45 "cups".  And then realized that 45 is neither even nor divisible by 16...
  

By the way, it is rather easy to divide a circle with a compass into 48 equal parts, or 3 x 2^n parts for any n.  Dividing it into 5 x 2^n or 5 x 3 x 2^n parts would be much harder (I knew how to do it but I am not sure I would remember now).  Dividing it into 45 parts, or 9 x m for any m, is impossible.  One would have to divide into 15 arcs with the compass and then divide each arc into three by hand or by trial and error.

All the best, --stolfi

[eggyk]

A lot of the illustrations in the VMS are simply dumb. The animals in Q13 next to the mermaid are totally anatomically wrong. Basic things like proportions, body types of animals are comparable to drawings made by absolute art-less amateur peasants, if not worse.

In fact, I'd even go as far as to say the VMS "artist" hasn't the slightest clue what a lobster even is. The same goes for other animals. What kind of person has never seen dogs, horses and lobsters?

Half of all of the artists in the 12th-15th century would fit in that category then, to be honest. The illustrations in the VMS are reasonably consistent actually, and of "ok" quality with proportions. They are nowhere near the worst out there. 

The cancer depictions are completely fine, other than where they drew the hind legs. They're not even in the bottom half of what I would consider a bad lobster example. Look at these!

           

And there are also examples of there being hind legs where they were drawn in the VMS: 
               

As for the rest of the animals, compared to some more professional drawings yes they are worse in quality. But they are not the worst out there by any stretch of the imagination. I think the average ability to draw anatomically correct animals in tiny, miniature sized drawings is being overstated. 

In fact, the size is often completely forgotten in these cases. The voynich illustrations are genuinely tiny, and they get compared to examples 5x bigger in size! Of course they will be lower quality if not done by a talented professional.

[rikforto]

Another thing I think we can assume to a degree is that those examples of manuscripts deemed most valuable and most worthy of preserving to the digital age are those which were higher quality, in part because of drawings. So not only do I agree we are overestimating the artistic ability in the surviving corpus, I would expect that represents a ceiling on the level of artistic talent we might infer for the pool of medieval manuscript artists

Given the estimated "only 5% of manuscripts survived" figure I see floating around, for every Voynich "dear God, why does that scorpion look like a lizard!?" horror there were 19 more artistic abominations that we will never get to see.

[ololololo]

Both are rather laborious to encode and decode.

Typically, arguments against ciphers focus on the fact that, supposedly, 200 pages would be very difficult to decipher.
But I’ll draw your attention to something slightly different:
1). You mean that the encryption scheme is very complex, but it will be complex if you want to encrypt a book very quickly. You won’t be able to write a 200‑page book in a week, even in unencrypted form. Why couldn’t the author simply come in the evenings, take the parchment, and gradually encrypt the text? Moreover, if one person was doing the encrypting (and I don’t think that’s the case), they could have done it in a fairly measured and lengthy manner. And in their old age, let our beloved “silly scribes” copy their drafts onto parchment.
This is regarding the encryption.

2). VMS is hardly some kind of fairy tale. It’s unlikely that the text spanning 200 pages is coherent. The maximum is Quire 13 and Pharma, where the text is clearly connected in terms of meaning, but only within the section. That is, if you need to extract some information from this book, you can use the key to open a couple of the necessary pages and decipher them. If the difficulty for you lies in the number of pages, then there isn’t any here. In Quire 13 there are 20 pages. I don’t think that even if it were a book cipher, it would be that difficult.
This is regarding the decryption.

Of course, these are all assumptions, and I can’t prove it... But at the very least, this is a reminder that sometimes you need to doubt. Even your own words