The Voynich Ninja

(22-07-2018, 12:51 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(21-07-2018, 10:24 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.I don't think VMS is written Classical Latin or classical Medieval Latin, and there is no need to kick at open door and compare them directly. Rather VMS is written in some form which mixes traits of Medieval, Vulgar and spoken Latin with artificial endings invented by the author.

Thank you farmerjohn, I see your point.  You understand your fake-Latin cannot stand a comparison with true Latin or Romance languages. Your only choice is discarding your theory or discarding the evidence of the entire tradition of the Latin language and all its descendants, and you are happy with the second option. A situation well described by Rene You are not allowed to view links. Register or Login to view.:

Quote:The worst case (also not rare) is when a theory is used to decide which evidence is reliable, and which is not.

A gaping abyss rather than "an open door", in my opinion.

---

(21-07-2018, 10:24 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.If it walks like a duck and quacks like a duck I’ll call it a duck

Appropriate enough: once you decide your results cannot be directly compared with actual evidence, you are left with pure flatus vocis - "quack quack quack", if you prefer.

My actual position is actually never mix problems and terminology
While mapping Voynichese to any real language is failing at the moment, is absolutely eligible to to try another approach and relax some conditions for example.
I respect you and Rene and many others greatly, but to solve well-known problem and to satisfy you are slightly different tasks, although they have a lot in common.
As for terminology, that's the worst matter for dispute. If some language has roots and suffixes from Latin, pronunciation from Latin then I believe it's acceptable to use word "Latin" when naming this animal. Other may have different opinions. That's absolutely ok.

(22-07-2018, 08:08 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.I respect you and Rene and many others greatly, but to solve well-known problem and to satisfy you are slightly different tasks, although they have a lot in common.

The relationship between a solution and consensus by other people depends on what you mean by "solution".

Effective solutions (like, say, Champollion's deciphering of the hieroglyphs) are supported by direct comparison with actual evidence. This allows others to check if the solution is correct. If it does a good job at explaining a significant part of the available evidence, in the end, most of the others will adopt it and build upon it (but, around the world, there will always be some who disagree).
In this sense, solving a problem and obtaining the consensus of other researchers are strongly correlated.
In this sense, the VMS hasn't been solved yet, even if several small steps in explaining some of the evidence have been taken.

Fake solutions like those by Michael Hoffmann, luiscrassus, Patrick Lockerby, Nicholas Gibbs or yourself (all more or less in the crowded fake-Latin arena) can only fool their creators and others who do not care about direct comparison with actual evidence. Gibbs for instance was taken seriously by several people, thanks to an article on the Times Literary Supplement; in the end You are not allowed to view links. Register or Login to view. had to point out that his Latin did not make sense. In these cases, a convergence between the views of the delusional creator and other researchers is impossible. By giving up discussing evidence, fake-solvers also give up any constructive discussion with others: they are only interested in their own fictional worlds. As they disregard actual evidence, they should also disregard the opinion of others: I guess many follow such a consistent behaviour but of course they are invisible, so it is impossible to tell.
In this sense, the VMS is being "solved" about once a month.

(23-07-2018, 09:02 AM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(22-07-2018, 08:08 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.I respect you and Rene and many others greatly, but to solve well-known problem and to satisfy you are slightly different tasks, although they have a lot in common.

The relationship between a solution and consensus by other people depends on what you mean by "solution".

Effective solutions (like, say, Champollion's deciphering of the hieroglyphs) are supported by direct comparison with actual evidence. This allows others to check if the solution is correct. If it does a good job at explaining a significant part of the available evidence, in the end, most of the others will adopt it and build upon it (but, around the world, there will always be some who disagree).
In this sense, solving a problem and obtaining the consensus of other researchers are strongly correlated.
In this sense, the VMS hasn't been solved yet, even if several small steps in explaining some of the evidence have been taken.

Fake solutions like those by Michael Hoffmann, luiscrassus, Patrick Lockerby, Nicholas Gibbs or yourself (all more or less in the crowded fake-Latin arena) can only fool their creators and others who do not care about direct comparison with actual evidence. Gibbs for instance was taken seriously by several people, thanks to an article on the Times Literary Supplement; in the end You are not allowed to view links. Register or Login to view. had to point out that his Latin did not make sense. In these cases, a convergence between the views of the delusional creator and other researchers is impossible. By giving up discussing evidence, fake-solvers also give up any constructive discussion with others: they are only interested in their own fictional worlds. As they disregard actual evidence, they should also disregard the opinion of others: I guess many follow such a consistent behaviour but of course they are invisible, so it is impossible to tell.
In this sense, the VMS is being "solved" about once a month.

Dear MarcoP, everything you wrote is very very (very) interesting and I hope you will not edit or delete this post so we can reread it in future.

But still I have a question for you. If one decides to publish his solution, which is mapping to some language, and reveals the dictionary, translations, word formation rules, phonetical rules, grammatical rules and with all that it's possible to translate the whole VMS - is it enough for you? Or you need something else? Or on the contrary something from this list you don't need? Precise list will be appreciated.

(23-07-2018, 09:29 AM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Dear MarcoP, everything you wrote is very very (very) interesting and I hope you will not edit or delete this post so we can reread it in future.

Thank you for your kind words. You have quoted my post, so this invaluable gem will be available to posterity in its original form even if I edit it.

(23-07-2018, 09:29 AM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.But still I have a question for you. If one decides to publish his solution, which is mapping to some language, and reveals the dictionary, translations, word formation rules, phonetical rules, grammatical rules and with all that it's possible to translate the whole VMS - is it enough for you? Or you need something else? Or on the contrary something from this list you don't need? Precise list will be appreciated.

Of course I am interested in these subjects, as long as they are discussed on the basis of verifiable evidence from the corpus of the proposed target language.  You have made clear that you are not interested in this, and your work is clearly based on this peculiar methodological choice. I will also add that I am more interested in "analysis" (sorry, David) than in "revelation".

As an example of the analytical approach I appreciate, some researchers have proposed tests that can help validate the results of a linguistic interpretation. You mentioned two of them here:

(21-07-2018, 12:38 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Two tests for solutions were mentioned in other thread. I don't find the one of Klaus Schmeh to be very usefull, but the other one (translating most frequent words, mentioned by ReneZ) is much more interesting.

You dismissed Klaus' proposal without a single word of comment. What a pity! Why don't you try it and see what happens?

You say Rene's test is "interesting" (like my post you kindly quoted above) and totally ignore its results. I wonder what your "interest" is, but of course this again involves discussing evidence, so it's understandable that you don't care.

I attach a table where I have applied Rene's test to three different Latin texts (excerpts from Virgil, the Vulgate, and Mattioli's 1554 book), a medieval Italian poem (Dante's Divina Commedia) and your own fake Latin. For each sample, I have only considered the 10 most frequent words. The benchmark I used (first column) is a list of the 10,000 most common Latin words I found You are not allowed to view links. Register or Login to view.. Honestly, I don't know on which works the list is based, but empirically it results to be consistent with actual texts.

For each word, I count the position in the target lists. Words that are so rare that they don't appear in the list get a distance value of 10001.
Taking the average position of each word in the target list results in a "distance" measure in the range "5.5-10001".  A good match will have a low value, a poor match a high value.
As expected, the three Latin texts get low values. Virgil has the highest distance (25) so the target list must be based on late Latin.
Italian has a 100 times higher value: 2678: this gives us a measure for a language which is definitely different from Latin, but not totally unrelated. We can safely assume that all Latin texts will produce a value considerably lower than this.

Your fake-Latin has a distance which is more than double as high as Italian: 6088. Obviously, the fake is not Latin. Possibly, the figure is made clearer by considering the last column: according to this measure, Shakespeare's English is closer to Latin than the fake!

(23-07-2018, 12:39 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(23-07-2018, 09:29 AM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Dear MarcoP, everything you wrote is very very (very) interesting and I hope you will not edit or delete this post so we can reread it in future.

Thank you for your kind words. You have quoted my post, so this invaluable gem will be available to posterity in its original form even if I edit it.

(23-07-2018, 09:29 AM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.But still I have a question for you. If one decides to publish his solution, which is mapping to some language, and reveals the dictionary, translations, word formation rules, phonetical rules, grammatical rules and with all that it's possible to translate the whole VMS - is it enough for you? Or you need something else? Or on the contrary something from this list you don't need? Precise list will be appreciated.

Of course I am interested in these subjects, as long as they are discussed on the basis of verifiable evidence from the corpus of the proposed target language.  You have made clear that you are not interested in this, and your work is clearly based on this peculiar methodological choice. I will also add that I am more interested in "analysis" (sorry, David) than in "revelation".

As an example of the analytical approach I appreciate, some researchers have proposed tests that can help validate the results of a linguistic interpretation. You mentioned two of them here:

(21-07-2018, 12:38 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Two tests for solutions were mentioned in other thread. I don't find the one of Klaus Schmeh to be very usefull, but the other one (translating most frequent words, mentioned by ReneZ) is much more interesting.

You dismissed Klaus' proposal without a single word of comment. What a pity! Why don't you try it and see what happens?

You say Rene's test is "interesting" (like my post you kindly quoted above) and totally ignore its results. I wonder what your "interest" is, but of course this again involves discussing evidence, so it's understandable that you don't care.

I attach a table where I have applied Rene's test to three different Latin texts (excerpts from Virgil, the Vulgate, and Mattioli's 1554 book), a medieval Italian poem (Dante's Divina Commedia) and your own fake Latin. For each sample, I have only considered the 10 most frequent words. The benchmark I used (first column) is a list of the 10,000 most common Latin words I found You are not allowed to view links. Register or Login to view.. Honestly, I don't know on which works the list is based, but empirically it results to be consistent with actual texts.

For each word, I count the position in the target lists. Words that are so rare that they don't appear in the list get a distance value of 10001.
Taking the average position of each word in the target list results in a "distance" measure in the range "5.5-10001".  A good match will have a low value, a poor match a high value.
As expected, the three Latin texts get low values. Virgil has the highest distance (25) so the target list must be based on late Latin.
Italian has a 100 times higher value: 2678: this gives us a measure for a language which is definitely different from Latin, but not totally unrelated. We can safely assume that all Latin texts will produce a value considerably lower than this.

Your fake-Latin has a distance which is more than double as high as Italian: 6088. Obviously, the fake is not Latin. Possibly, the figure is made clearer by considering the last column: according to this measure, Shakespeare's English is closer to Latin than the fake!

MarcoP, please allow me not to address points I've already addressed. I don't want to go on the second lap.

On new points:

I still wait for the list

Rene's test deals with words, Schmeh's deals with phrases. At the moment Rene's one is more suitable for me, for more developed theory one can use Schmeh's one.

Also subtype of Rene's test is inverse of Rene's test - proposed translations of frequent Voynich words must also be frequent. And this variant is much more important than the direct one. IMHO, of course

(23-07-2018, 03:13 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Also subtype of Rene's test is inverse of Rene's test - proposed translations of frequent Voynich words must also be frequent. And this variant is much more important than the direct one. IMHO, of course

I am not sure I understand. Could you please share a table with actual numbers like I did?

(23-07-2018, 03:33 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(23-07-2018, 03:13 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.Also subtype of Rene's test is inverse of Rene's test - proposed translations of frequent Voynich words must also be frequent. And this variant is much more important than the direct one. IMHO, of course

I am not sure I understand. Could you please share a table with actual numbers like I did?

I just mean that if we map Voynichese to language, then frequent words of language should be frequent in Voynichese AND frequent in Voynichese should be also frequent in language.

By the way obviously any other property can be tested in both directions

(23-07-2018, 04:22 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.I just mean that if we map Voynichese to language, then frequent words of language should be frequent in Voynichese AND frequent in Voynichese should be also frequent in language.

By the way obviously any other property can be tested in both directions

I see, thank you. Your fake-Latin fails in both directions:

* As I showed above, the top 10 most frequent fake-Latin words (partellus & C) averagely are at position 6088 (there are more than 6087 Latin words that are more common) instead of the expected 50 or less. The most frequent ("illi") is only at position 128 in true Latin.

* None of the top 10, and only two of the top 100 Latin words appear in your fake top 20 words ("ex", #13 in the target list, and "iam" #44).

Where are all the frequent Latin words listed You are not allowed to view links. Register or Login to view.? Clearly, this isn't Latin, right?
But you should consider doing these checks yourself, I know it's hard work, but they will be much more instructive that way.

As for the list you need, I don't have time to work on it right now, I am sorry. I'll be happy to contribute in the future. 
I think you should start a new thread for this list of requirements: it will not be easy to define and it could be useful to other researchers as well. We could collect a set of guidelines that could make a candidate solution easier to present and to evaluate (this could include a few simple tests like the one proposed by Rene).

(23-07-2018, 05:12 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(23-07-2018, 04:22 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.I just mean that if we map Voynichese to language, then frequent words of language should be frequent in Voynichese AND frequent in Voynichese should be also frequent in language.

By the way obviously any other property can be tested in both directions

I see, thank you. Your fake-Latin fails in both directions:

* As I showed above, the top 10 most frequent fake-Latin words (partellus & C) averagely are at position 6088 (there are more than 6087 Latin words that are more common) instead of the expected 50 or less. The most frequent ("illi") is only at position 128 in true Latin.

* None of the top 10, and only two of the top 100 Latin words appear in your fake top 20 words ("ex", #13 in the target list, and "iam" #44).

Where are all the frequent Latin words listed You are not allowed to view links. Register or Login to view.? Clearly, this isn't Latin, right?
But you should consider doing these checks yourself, I know it's hard work, but they will be much more instructive that way.

As for the list you need, I don't have time to work on it right now, I am sorry. I'll be happy to contribute in the future. 
I think you should start a new thread for this list of requirements: it will not be easy to define and it could be useful to other researchers as well. We could collect a set of guidelines that could make a candidate solution easier to present and to evaluate (this could include a few simple tests like the one proposed by Rene).

That's clearly not Classic Latin and not Medieval Latin. So to compare them we need transform lists.
For example if we compare most frequent words of English and Russian (my native language), we immediately remove articles like the and a and auxiliary verbs from one list and merge different forms of the word in another list.
For Classic vs "fake" we need similar amount of work.

But ok, imagine we have done this. Top 10 Classic is among top 30 Voynichese and vice versa.
Does that proof something? No
Does that disproof something? No
Does that help to correct errors? No

In situation when you lack almost everything any action without result is unaffordable waste of resources. And when you know the situation from inside including all issues and weak points you naturally select different strategy than the person who looks from outside.

(23-07-2018, 07:06 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.That's clearly not Classic Latin and not Medieval Latin. So to compare them we need transform lists.
For example if we compare most frequent words of English and Russian (my native language), we immediately remove articles like the and a and auxiliary verbs from one list and merge different forms of the word in another list.
For Classic vs "fake" we need similar amount of work.

But ok, imagine we have done this. Top 10 Classic is among top 30 Voynichese and vice versa.
Does that proof something? No
Does that disproof something? No
Does that help to correct errors? No

Hi Farmerjohn,
from an analytical point of view, Rene's test counts as evidence of how a "solution" compares with the target language. I don't know if it is a "proof", but it certainly is a relevant quality index. From this evidence-based point of view, your current solution is clearly dismissed. A formalized transformation resulting in matching the top 10 Latin words with the top 30 fake words and vice-versa would be a huge step forward: your fake could look much less fake, at least from the lexical point of view. Obviously, this would imply the total replacement of bogus words like "cartellus", "partellus", "eare" etc with words from the actual top 30 Latin words. Doing this would be "correcting errors".
Honestly, I don't think there is any possibility of matching Voynichese and Latin (different word-length histograms, different entropy values, reduplication frequent in the VMS and absent in Latin, preponderance of variable prefixes in the VMS vs variable suffixes in Latin etc). If you have a concern with economizing resources, Latin is a very poor choice.

I still think that a thread about "a list of requirements" could be a good idea: if you start it, I will contribute my opinion in due time.

I am grateful for this conversation. I now have a simple tool based on Rene's idea that I can run to evaluate any solution claiming to be based on Latin. I guess it could come handy in the future.