The Voynich Ninja

(09-09-2016, 11:04 PM)ThomasCoon Wrote: You are not allowed to view links. Register or Login to view.

(09-09-2016, 03:46 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.

(09-09-2016, 02:43 PM)ThomasCoon Wrote: You are not allowed to view links. Register or Login to view.1) If you are searching for a natural language for Voynichese and believe the VMS spaces are true, you are almost certainly limited to agglutinative and polysynthetic languages because of the "morphology" of Voynich words (see the <qot> forms). This eliminates most of the languages in Europe.

I see your point. This seems to be consistent with where some linguistic research about the ms is going.

Interesting, Marco; could you tell me which researchers think this?

Hi Thomas,
I was mainly thinking of Stephen Bax, who also is of the opinion that European languages are not good candidates for Voynichese; he also You are not allowed to view links. Register or Login to view. that a phenomenon like the common EVA:o- prefix might be an agglutinating article analogue to the Arabic al- and that the prefix EVA:y- might function as an agglutinating disjunction.

But for strictly polysynthetic languages, one also thinks of Tucker and Talbert's Nahuatl theory (of which I know very little).

Hey all, I've been trying to see if I could prove that spaces (as written) in the VMS are fake. I found some things which may indicate patterns across words, but I submit the images for your review:

<otchordy> split two different ways on the same page?


<otchordy> and <opchordy> in the text as 1, 2, or 3 words:

And then there is this beauty, where <kydaiin> repeats, but the combinations of <ch-ky> and <sh-ky> are also interesting:



Hope this might catch your interest!

(15-09-2016, 02:28 AM)ThomasCoon Wrote: You are not allowed to view links. Register or Login to view.Hey all, I've been trying to see if I could prove that spaces (as written) in the VMS are fake. I found some things which may indicate patterns across words, but I submit the images for your review:

<otchordy> split two different ways on the same page?


<otchordy> and <opchordy> in the text as 1, 2, or 3 words:

And then there is this beauty, where <kydaiin> repeats, but the combinations of <ch-ky> and <sh-ky> are also interesting:



Hope this might catch your interest!

I've also been looking at the spaces, because it can potentially answer many questions.

For example, I asked,

• if spaces have been added (or removed), does that mean "unique" words are not necessarily unique?
  
• if spaces have been removed, does that mean the gallows characters might be pilcrows/capitula?
  
• if spaces have been added, does that mean the labels map more widely to the main text than currently appears?
  
• if spaces have been added/removed, do the changes in units that appear to behave like prefixes/suffixes have a different function?
  

I had looked at the EVA font and didn't agree with some aspects, so I created my own. Then I glanced at the first three pages of the Takahashi transcription (this was years ago so I don't know if I remember the name correctly), didn't agree with that either and created my own transcription of the entire manuscript. I also created a desktop-publishing version of the VMS font that looks like Voynich glyphs but is easier to read than regular Voynichese.

Then I analyzed the text and broke it into its component parts.

Some time ago, I created a list of what I believe to be atomic units (root words, prefixes, suffixes and those which may be markers/modifiers).

• I created a color-coded concordance of every vord in the manuscript (this took months of dedicated effort and was only the beginning).
  
• I located the unique words, broke them into units and followed those units to see whether the components were found elsewhere with spaces. I also organized them so I could see if unique-word units were manipulated in some consistent way and whether their position relative to illustrations and other words changed significantly when they were deconstructed. In part, I was trying to determine whether spaces could be believed and, if not, whether they were arbitrary or contrived.
  
•  I then started working on word pairs and word triples to see if they behaved in the same way or yielded different patterns. This is a massive task but is necessary if the spaces are not always real.
  

Before you ask me to post examples, note that this is a huge project. I have over 2,000 pages of notes describing what happens and I haven't even noted it all down yet or "pretty formatted" all the charts. Making lists of word-token patterns based on respecting the spaces (a worthwhile examination in its own right) is trivial compared to describing what happens when spaces are added and subtracted. It's also trivial compared to extrapolating the meaning of the patterns.

Most of my notes are based on describing the significance of the patterns rather than listing them (both together would be thousands of pages longer). The lists can be generated algorithmically, so I've left that on my "to do" list. As for significance, I think I have a fairly good idea of what some of the dense-text pages (the ones without illustrations) are about. I also think I know how the starred text pages relate to the rest. I am still somewhat unsure about the "pool" pages but I have some suspicions (too much to do, not enough time). In general, I am able to recognize some of the VMS vord groups and predict where they will show up and approximately how often. Even this took months to achieve.

But, to give a few examples that illustrate the scope of this line of research, consider these (and note that these are ONLY examples, there's no way to summarize 2,000+ pages of notes on a forum post or even in a paper without over-simplifying it):

• oPnccol appears as unique to the star section, but appears as nccol in a number of specific places. Similarly, Hñcoßal is unique to a plant page but ñcoßal and also ñco ßal are found elsewhere (sorry, I don't use EVA, I have my own system). Note that both are preceded by gallows characters. HñoIcoßg and Pncoßng also behave this way. They are at the beginning of paragraphs, and are unique words until you remove the gallows character. This adds data to the question of whether gallows might be pilcrows/capitula.
  
• Most unique words break down into two components that can function independently, but some break into three. Breaking them into two components does not mean those components are frequent (as one might expect if they were linguistic syllables).
  
• When a unique word does not break into atomic units that function independently, most of the time it's because a single character has been added at the end. Remove this and they usually map to another section. One has to ask if it's a null (or a modifier such as we might add "-s" to create plurals).
  
• When a unique vord is short, adding a prefix or suffix will usually relate it to vords elsewhere.
  
• Vords like PoxaRaR, which is always at the beginnings of paragraphs, break down into 1) a unit that is usually at the beginning of paragraphs and 2) one that has a limited distribution in related sections.
  
• Vords beginning with 4o tend to cluster in two specific kinds of pages.
  
• Vords that behave as suffixes will frequently appear after many of the same units with a space in between. In other words, if you have something like abcdefg and "fg" is a frequent suffix, then it's also common to find abcde fg with a space in between. This suggests that some spaces may be contrived but even if they are, it doesn't mean all units necessarily behave this way.
  
• Just as there's a high level of clustering of 4o words in certain sections, there is a strong clustering of oH words in the zodiac pages and where they lead depends on what follows the prefix. This, combined with similar examples, demonstrates that the text is not random and is not created with a grille to make it look less random. It is systematic and when you get to know the system well enough, you can predict where to find certain vords.
  
• The length of a vord does not always predict its frequency. You might think longer words would be less frequent, short words might be moreso (as in natural-language patterns of "a" "in" "of" "the" "is", etc.) but some short words are quite infrequent and there's a certain group of long words that are fairly frequent.
  
• The distribution of some of the vords suggests that some of the drawings might not be what they seem at first glance. They also suggest that there are connections between separate sections that make the entire manuscript a somewhat cohesive effort (rather than a bunch of separate pamphlets on different topics strung together). I haven't had time to check if all the sections behave this way, but I've checked many of them.
  
• Even though there is strong evidence of digraphs, many of the prefixes and suffixes are single characters.
  

Here is another observation that argues against random or grille-generated text:

There are two kinds of "unique" words:

• those that can only be found in one place (and which can usually be broken into independent units as noted above, which means they may not be unique), and
  
• those that map a page to the starred text (to one or multiple starred pages) and nowhere else. One way to interpret this is that the starred text may be summary information that ties together the preceding pages. I mean "summary" in the broad sense. If the plants were medicinal (for example) and the starred text were prescriptions, one would expect this kind of dynamic but there are other possibilities, as well.
  

A large number of plant pages have vords that map only to the starred text. When they do, there are usually one to three of these vords and they are usually spaced out (not adjacent). The zodiac pages have them with similar frequency but the small-plant pages have noticeably fewer in proportion to the amount of text.


Thomas, this is a significantly condensed version because there's no way to explain any of this with so few words and I had hoped to write up the long version , with charts, arrows, balloons, and everything else that's required to explain it properly (along with some of my ideas of what some of the vords mean), but there's a lot of interest in this now so there you go. I can't sit on it forever.

Wow, JKP, thank you! That sounds like a massive undertaking, and very impressive! I may have seen one or two things dimly and "through a veil" but you see everything clearly and in daylight . I found your post very enlightening, and I see you have answers to many of my questions (ex. how paragraph-initial gallows may work, why and where spaces occur, etc.)

Your quote here:

Quote:Vords that behave as suffixes will frequently appear after many of the same units with a space in between. In other words, if you have something like abcdefg and "fg" is a frequent suffix, then it's also common to find abcde fg with a space in between. This suggests that some spaces may be contrived but even if they are, it doesn't mean all units necessarily behave this way.

I agree. I think I have seen this with -dy. If you say that some (but not all) spaces can be contrived, I see how that can be the case also.

Can I ask, when you say "Most unique words break down into two components that can function independently, but some break into three," do you mean the type of thing that I saw in my "Breakdown" thread: independent units like <qo, or, ol>? Or something different?

I find it highly interesting that you can predict where certain vords will be. And the fact that many words on plant pages map only to the starred text may be very significant. You probably know this, but German medieval doctors believed that both plants and starlight affected the human body. Paracelsus believed that light from certain stars brought diseases and poisons, and he told patients to sit under different stars to be cured. However, he also used plants medicinally, and so did many other medieval doctors, so the link between these two sections would definitely make sense.

Quote:Thomas, this is a significantly condensed version because there's no way to explain any of this with so few words and I had hoped to write up the long version , with charts, arrows, balloons, and everything else that's required to explain it properly (along with some of my ideas of what some of the vords mean), but there's a lot of interest in this now so there you go. I can't sit on it forever.

Yeah - it was clear from your other posts that you had many insights into how the text worked. Thank you very much for the explanation - I'll definitely be excited to see where you take things from here, and how I might apply your understanding. Lately I've been thinking about putting my Voynich work "in the back seat" and coming back to it later, because it has grown to an obsession and a time-sink, but I think what you're doing is very fascinating and it never once crossed my mind to figure out vords through context like you are doing! Highly exciting!

ThomasCoon queried: Can I ask, when you say "Most unique words break down into two components that can function independently, but some break into three," do you mean the type of thing that I saw in my "Breakdown" thread: independent units like <qo, or, ol>? Or something different?

Many of the unique words are of medium or longer length, so the breakdown units are not always small. For example, abcdefpqrst might break into abcdef pqrst—which function independently. Sometimes (not always), those broken-down components will include some of the smaller units (usually ones that appear to be digraphs) but it's hard to know if they are intended to be deconstructed all the way.


To give examples in English...

• Some words are combinations of syllables. A word like effervescent uses the same suffix as luminescent or incandescent. In this case, the suffix "-ent" turns a verb into another grammatical unit. The fact that "cent" is at the end of all of them, and can be a word on its own (denoting a penny), is a coincidence. Note also that the word "incandescent" includes "can", which is also a word on its own, but is not related to the "can" in "incandescent". Thus, patterns that appear to be atomic units might not be related.
  
• In contrast, some words are combinations of other words like "railroad" or "grandmother". These break down into two words that not only function on their own, but retain the same general meaning when they are on their own and breaking them down further wouldn't make it any clearer.
  

In some cases, the unique words in the VMS appear to be of the latter kind and sometimes of the former. The only way one can actually see this (or make some good guesses about it) is to map every word in the manuscript.

After doing that, I noticed that some of the smaller components appear to behave differently and are ubiquitous, which suggests they might be modifiers or linguistic units (basic building blocks) as in the first example, and others appear to be stand-alone units (I'm still reluctant to call them words), as in the second.


One begins to appreciate why this hasn't been decoded using traditional methods of attack.



Thomas wrote: I find it highly interesting that you can predict where certain vords will be.

I didn't break out the wine bottle until I was able to predict twelve in a row. It got to the point where I could say, "This will go to the starred-text pages and nowhere else," or "This is a zodiac word," or "This page should have a word starting with __ somewhere in the middle of the second line," or "This will show up on the small-plant pages, probably the leaf section," etc.


It wasn't possible to do this until about 40% of the words had been mapped and, even then, only in a very limited way. After about 60% had been mapped, it was easier to see the patterns, and to appreciate how much work went into laying down the text. But, I still haven't broken out the champagne because I only have an inkling of what some of them mean. Finding them and predicting them is a big step, but it's not a translation.


Thomas wrote: And the fact that many words on plant pages map only to the starred text may be very significant. You probably know this, but German medieval doctors believed that both plants and starlight affected the human body. Paracelsus believed that light from certain stars brought diseases and poisons, and he told patients to sit under different stars to be cured. However, he also used plants medicinally, and so did many other medieval doctors, so the link between these two sections would definitely make sense.


First I should clarify my terms. When I talk about star pages, I mean the cosmological section. When I speak of starred text, I mean the dense text at the and with "flower" stars to the left of the paragraphs.

But to return to elaborate on your point, plants were believed to be governed by certain stars/planets, which might account for the word-relation trail the flows through the zodiac/plant/star pages.



In terms of philosophy... I did the same thing most people do when I first encountered the VMS. I tried applying different languages. I'm familiar enough with western languages and some of the Asian languages (Korean, Chinese, Japanese) to approach it this way and noticed that ancient Mediterranean languages, especially the Semitic ones, a few old eastern European languages, and Asian languages like Boa, fit the structure of the VMS better than most western European languages. But somewhere along the way I realized this might be an unproductive way to approach it. Not only have many tried, but there are thousands of languages (including extinct ones) and it could take forever to figure out which one it is, especially if the spaces (or some of them) are fake.

So I took a different road. I decided, why not just learn Voynichese? If it does turn out to be a natural language, it will become evident. If it's a constructed language, this will be evident also. The advantage of learning Voynichese, rather than cherry-picking languages and trying to impose them on the text, is that it encompasses both possibilities (and a few others) within the same methodology.

JKP
I have a remark on the likelihood of the use of Cardan grille. I have worked in practice with the lattice, so evaluate this probability of 0.01%. The amount of work from encryption to decryption process in this case, 3-fold increase (+ tablets when using wax 4-fold increase). This applies to the the increase in expensive consumables.
As the line spacing in the manuscript - floats (not permanent - no linovki empty feather (without ink)). For encrypting, you must to have a a duplicate an entire of the encrypted text on a sheet of lined notebook with the drawn contours of the drawings!
With by diagrams - circles particular problem.
 Then you need to to make a the manuscript in its present form. Thus the volume information drop (maximum) to 50% of characters (using numbering reading order of cells grating - to 75%).
To decrypt the need to rewrite existing manuscript on lined pages, but not just withstanding line spacing, but also intervals (programmed) between characters.
It is cessary find the anchor label of grating .
This would dare is only the crazy!
 The text complicated for Google translator. Duplicated the text in Russian.

У меня есть одно замечание по поводу вероятности использования решетки Кардано. Я работал на практике с решеткой, поэтому оцениваю  такую вероятность 0,01%. Объем работы от процесса зашифровки до расшифровки в нашем случае увеличится в 3 раза ( + при применении восковых табличек увеличение в 4 раза). Это касается и увеличение дорогих расходных материалов.
Учитывая то,  что межстрочное расстояние в манускрипте – плавает (не постоянное – нет линовки пустым пером (без чернил)). Для зашифровки, необходимо иметь дубликат всего  зашифрованного текста на разлинованном  листе с нарисованными контурами рисунков!
С диаграммами – кругами особая проблема.
Затем необходимо изготовить манускрипт в существующем виде.  При этом объем информации упадет (максимально) до 50% символов (при использовании, нумерации порядка прочтения ячеек решетки до 75%).

Для расшифровки  необходимо переписать существующий  манускрипт на линованных листах,  но уже  выдерживая не только межстрочные расстояния, но и интервалы (запрограммированные) между символами.
Необходимо найти метку привязки решетки.
На это отважится только сумасшедший

Wladimir, very interesting insights on your work with a grille.

(15-09-2016, 06:38 PM)-JKP- Wrote: You are not allowed to view links. Register or Login to view.Thomas wrote: I find it highly interesting that you can predict where certain vords will be.

I didn't break out the wine bottle until I was able to predict twelve in a row. It got to the point where I could say, "This will go to the starred-text pages and nowhere else," or "This is a zodiac word," or "This page should have a word starting with __ somewhere in the middle of the second line," or "This will show up on the small-plant pages, probably the leaf section," etc.

It wasn't possible to do this until about 40% of the words had been mapped and, even then, only in a very limited way. After about 60% had been mapped, it was easier to see the patterns, and to appreciate how much work went into laying down the text. But, I still haven't broken out the champagne because I only have an inkling of what some of them mean. Finding them and predicting them is a big step, but it's not a translation.

Wow! That's very enlightening! It sounds like you will be able to break out that champagne before too long - maybe a next step would be to identify parts of speech (nouns, verbs, etc.) if possible, and then see which properties all "verbs" have in common? Whatever the case, this is fascinating work  

Quote:In terms of philosophy... I did the same thing most people do when I first encountered the VMS. I tried applying different languages. I'm familiar enough with western languages and some of the Asian languages (Korean, Chinese, Japanese) to approach it this way and noticed that ancient Mediterranean languages, especially the Semitic ones, a few old eastern European languages, and Asian languages like Boa, fit the structure of the VMS better than most western European languages. But somewhere along the way I realized this might be an unproductive way to approach it. Not only have many tried, but there are thousands of languages (including extinct ones) and it could take forever to figure out which one it is, especially if the spaces (or some of them) are fake.

So I took a different road. I decided, why not just learn Voynichese? If it does turn out to be a natural language, it will become evident. If it's a constructed language, this will be evident also. The advantage of learning Voynichese, rather than cherry-picking languages and trying to impose them on the text, is that it encompasses both possibilities (and a few others) within the same methodology.

Brilliant idea to learn Voynichese, and definitely a unique approach. The thought to do that never crossed my mind - I started by making no assumptions about the text and just copied it, page by page, waiting for realizations to emerge ("this prefix keeps repeating" - "these letters always appear together" - etc.). Both of our approaches are grounded on observation and analysis, it seems!

Hello everyone!

I have been studying VM for some time and have written a short paper recently. It can be found You are not allowed to view links. Register or Login to view. or in the attachment. The paper proposes the key to manuscript and gives translation for several words. Also some thoughts about "EVA-o in the beginning" are presented, probably not unique ones.

If the work has any sense I will be happy

[attachment=621]

Hi farmerjohn and welcome to the forum!

Please keep the discussion on topic, this thread is dedicated to the text analysis by ThomasCoon. Please feel free to open a new thread dedicated to your own analysis (if you so wish).