The Voynich Ninja

In discussing a Voynich prize the issue of handling a large number of false decipherment "solutions" was raised.

I get the impression, though I haven't looked in depth at many false solutions, that they have a lot of common and so very similar flaws. Whether the theory is proto-turkic or proto-romance or some kind of vulgar latin it seems they have a "degree of freedom" problem such that there is sufficient flexibility in interpretation that a theory can be made to fit. It also appears that the process by which the theories were created has a lot in common.

It is my opinion that the real solution will have very little in common with these kinds of theories both in the method by which it was arrived at and the structure and nature of the actual solution.

However in order to reduce the time taken investigating false solutions it could be worth classifying them and so when a new theory is presented it can be classified as for example a "Cheshire solution" or other type. One can then point to the flawed method by which the solution was constructed and the flawed widely open to interpretation, inherently highly flexible solution  structure.

The degrees of freedom allowed in such solutions that we even see in the anagram "Artificial Intelligence" theory should be a warning that "there is something rotten in the state of Denmark"(apologies to any Danes; it's a Shakespeare quote) With sufficient degrees of freedom in interpretation then any theory will fit. As an absurd example if I postulate that every Voynich symbol can represent any letter in the alphabet then I can interpret the text to read whatever I wish. Whilst that is certainly an extreme example it appears to me that so many false solutions have such a large degree of freedom.

Managing these false solutions may help in the construction of standard tests for these theories.

So I see the handling of a large number of false solutions as unlikely to be problematic.

It is worth mentioning, I think, that often the method by which these solutions were arrived at is not explicitly stated by the author and so must be surmised. However I think it is not difficult to guess.

It is true that even when two theories postulate the Voynich is written in "vulgar latin" whilst their precise solutions might be quite different structurally they could be very similar. Given the huge range of possible interpretations it seems to me very unlikely that two solutions will be identical, and frankly if they are I would think they deserve serious consideration, but they can be classified as an identical type of solution if structurally they are the same.

I am specifically referring to decipherment/translation theories here in this thread. Another thread->

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

seems to be much broader in scope.

Frankly if someone's theory is that the Voynich was written by someone Jewish, absent of any specific translation, then they could be right. So I see far less reason to see that theory as problematic or in much need of "debunking". One might argue about the basis on which they justify that idea, but to dismiss the notion completely out of hand seems foolhardy to me. The same applies to many other fairly non-specific theories, or at least theories that could be said to lack much detail. Even when someone says the Voynich is a medieval hoax it is not an unreasonable idea, that needs "debunking", although it is certainly not my idea. By contrast proper translation/decipherment theories have a real degree of specificity.

The thread->

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


seems to be focused on whether it is worthwhile to argue with specific theories or not. That is not what I am addressing here. If someone poses a translation theory and nobody gives it a moment's consideration that seems unwise as there could be something there. Some theories rise to such a public profile that they have to be addressed. Therefore a method of addressing these theories becomes vital.

The reason this so far has been a one-man conversation is that this a truly problematic situation. There have been, over several years, a succession of proposed linguistic interpretations, proclaimed translations, suddenly flaring up - and then cooling off and fading away - trying to interpret the 'Voynichese' "language." And from these failures, how might we hope to 'get a jump on' any further 'potential' failures, as they will probably turn up in the prospective future? Do these failures fit into 'types' or classifications? And, if so, what types are they? And will the next proposed 'linguistic solution' fit into one of these prior categories? This sort of thing devolves into a fur ball of further problems.

As an alternative, perhaps, there could be the setting of a certain initial standard for any sort of "formal" consideration < whatever that is.> Two simple standards of interpretation have long been set forth: (1) that it is repeatable, and (2) that it makes sense. No reason that these standards should alter normal discussion, however, as they have yet to take effect.

The contest is to turn VMs glyphs into a known language. Not to present a 'translation', but to demonstrate a methodology. So many ways to go wrong, but is there, was there originally, a way to go right?

I don't think you can classify theories/false solutions, and so I'd disagree that there is a "Cheshire" type that is easily distinguishable from other theories.  

What we see is an array of symptoms that false solutions tend to display.  Every false solution has at least a combination of these.  Certain of the symptoms are seen in all theories; others may or may not be present, in varying combinations.  And as these combinations likely overlap, I don't think there is a way of making a clean distinction and saying "this theory has Combination A symptoms so is Type A theory", etc.   But it's probably true to say that the worse theories tend to have all or most of the symptoms; the better theories have fewer. 

As for the symptoms, a non-exhaustive list would include:  

1. Provides "translations" of isolated words, or at most sentences out of context
   
2. Uses phrases like "it cannot be a coincidence": incorrectly thinks the probability of finding a "matching" word in their target language is low; is unaware of how prevalent such coincidences are, and of confirmation bias
   
3. Has not read the arguments provided by other theories, and so repeats their mistakes
   
4. Searches a dictionary and picks the word they like the most to fit their system, even if it is a rare or odd word
   
5. No grammar:  any sentences are a "word salad", with words chosen for how well they fit the system rather than whether they fit into a grammatical sentence.  There is little or no attention paid to how well the "translation" matches up with the target language's grammar 
   
6. System has multiple degrees of freedom:  in a blind test, two users of the system would not arrive at the same translation.  The "translation" is not repeatable.  
   
7. Explains away difficulties by suggesting multiple languages are used, or says experts in the language are needed to provide a full translation.  Keeps adding excuses for why a longer translation cannot be provided and/or for why the system stops working when applied to longer phrases
   
8. Little or no thought given to how the author(s) would use this system in practice and whether the information is really recoverable 
   
9. Fails to explain certain or even all key Voynichese unlanguage-like behaviour (line start/end patterns, paragraph start/ends, letter positions, word break combinations, and this is not an exhaustive list).  Many don't even know of these patterns since they haven't done their research.
   
10. Poor or no methodology; no cataloguing and self-questioning of assumptions they are making, of any distortions or accommodations they make in order to produce a "translation"; does not act as their own sceptic, leaving that role to others
    
11. Reacts badly to criticism of the theory.  Spends all their time defending the theory rather than genuinely exploring whether criticisms are right.  Some may even turn abusive.
    
12. Finds the "solution" in a matter of days or weeks but never asks the question "If the solution was this obvious, why hasn't anyone found it before?"  Focuses instead about how they are the one person to finally find the solution, to pull the sword from the stone. 
    

Most of these overlap or cause each other, which is why I can't see how there could be a classification.  And ultimately I'd agree with R.Sale that it is methodology (no 10 in the list) that is probably the most important.  Bad/non-existent methodology is the one symptom that is present in every single wrong theory.  I also doubt that the "true solution" can be demonstrated without a strong methodology.  Not just because a strong methodology helps in the first place prevent you from going down the wrong track and ruining your chances of finding the true solution, but also because it will help reduce/explain away degrees of freedom and show that your solution is repeatable.

(05-12-2021, 10:57 PM)R. Sale Wrote: You are not allowed to view links. Register or Login to view.The contest is to turn VMs glyphs into a known language. Not to present a 'translation', but to demonstrate a methodology. So many ways to go wrong, but is there, was there originally, a way to go right?

I believe the first step in solving the VM is to find the sources for the alphabet in the VM, in particularly the shapes of the letters and ligatures, not to guess the letters by assuming that a certain word means something in a particular language, because that same word can mean something different in a different language. 
There is a way to guess some letters, like the Zandbergen and Landini did with 's' and 'b', which look like a mirror image of the Latin 's' and 'b'.  
Following this, I had found all, but EVA-q and four tall VM glyphs in the 15th century European manuscripts. To this point, my transliteration alphabet can be replicated. I designated the VM tall letters for the letters and letter combinations in the comparative manuscripts, which do not appear in the VM. If I was able to do that, others can do that too. I am talking about the shapes of the letters. It would be interesting how many words in different languages can be obtained with the letter-to-letter transcription. 


Without knowing the language, VM researchers have come to important observations about the VM prefixes and suffixes. Using the proper transliteration alphabet would enable to properly interpret the prefixes and endings. The variety of the same root words and different endings indicates the language is highly inflective.
The problem with highly inflective languages is that the words appearing in the VM, even if properly transcribed into Latin letters, cannot be found in any dictionary. 

It would be most logical that the person having right (or more right) solution would be able to translate the words that appear most frequently in the VM, and not the one single word from the floral page, since the names of the flowers might be in a foreign language.  The most frequently used words  would be the oldest and perhaps still in use in the VM language.

The linguistic solution would also have to agree with other facts, like fashion, zodiac symbols, relationship to Emperor Rudolf II etc.

(06-12-2021, 01:47 AM)tavie Wrote: You are not allowed to view links. Register or Login to view.I don't think you can classify theories/false solutions, and so I'd disagree that there is a "Cheshire" type that is easily distinguishable from other theories.  

What we see is an array of symptoms that false solutions tend to display.  Every false solution has at least a combination of these.  Certain of the symptoms are seen in all theories; others may or may not be present, in varying combinations.  And as these combinations likely overlap, I don't think there is a way of making a clean distinction and saying "this theory has Combination A symptoms so is Type A theory", etc.   But it's probably true to say that the worse theories tend to have all or most of the symptoms; the better theories have fewer.

I think you have made a good point as far as the difficulty of classifying theories.

As far as your list goes I find some of your items much more interesting than others, namely the following:

(06-12-2021, 01:47 AM)tavie Wrote: You are not allowed to view links. Register or Login to view.As for the symptoms, a non-exhaustive list would include:  

1. Provides "translations" of isolated words, or at most sentences out of context
   
2. Searches a dictionary and picks the word they like the most to fit their system, even if it is a rare or odd word
   
3. No grammar:  any sentences are a "word salad", with words chosen for how well they fit the system rather than whether they fit into a grammatical sentence.  There is little or no attention paid to how well the "translation" matches up with the target language's grammar 
   
4. System has multiple degrees of freedom:  in a blind test, two users of the system would not arrive at the same translation.  The "translation" is not repeatable.  
   
5. Explains away difficulties by suggesting multiple languages are used, or says experts in the language are needed to provide a full translation.  Keeps adding excuses for why a longer translation cannot be provided and/or for why the system stops working when applied to longer phrases
   
6. Little or no thought given to how the author(s) would use this system in practice and whether the information is really recoverable 
   
7. Fails to explain certain or even all key Voynichese unlanguage-like behaviour (line start/end patterns, paragraph start/ends, letter positions, word break combinations, and this is not an exhaustive list).  Many don't even know of these patterns since they haven't done their research.
   

I would add that I think that when you say "incorrectly thinks the probability of finding a "matching" word in their target language is low; is unaware of how prevalent such coincidences are, and of confirmation bias" that is a good observation, although I see no problem in general of using the phrase "it cannot be a concidence"

I would say that "no cataloguing of any distortions or accommodations they make in order to produce a 'translation'" seems a like a good point to me. The rest of that item seems more general and broader in its scope and so less interesting.



The following items I find much less interesting as they are much more general and not theory specific in the way that I am trying to get at:

[*]
Nevertheless I appreciate your points as some get much closer to the subject I was trying to address.

[*]