The Voynich Ninja

(13-04-2019, 10:03 AM)Koen G Wrote: You are not allowed to view links. Register or Login to view.But this is the problem: if you need illustrations to pattern-guess the solution, it's a one-way cipher. A large part of the VM text is without illustrations.

I think it depends on how much practice and experience one has reading it. None of us here has much practice and experience reading it: I am learning it myself, Koffee is beginning to learn it, and the same goes for anyone else here who is trying to learn how the method works.

But imagine, what if this were a writing system that got taught and learned and practiced regularly and daily and passed down from generation to generation? I am confident that if you taught this writing system to Greek schoolchildren starting in kindergarten, and they practiced reading it every day, then by the time they were adults, they could potentially be able to read even the Voynich MS pure text without any illustrations or context clues. 

To begin with, however, one learns to read with simple texts: repetitions of well-known prayers one has already spoken and heard hundreds or thousands of times, short simple statements with accompanying illustrations to guide one, etc. This is how children learn to read. When it comes to this new method to read an unfamiliar script, we here and now are all in the position of children learning how to read it. But if one had started learning it when one was 5 years old, and read it ever since, as a native speaker of the language, I believe after decades of practice one could read even a text like the Voynich MS.

The Arabic script without the dots is technically speaking a "one-way cipher" in cryptological terms. And yet many experienced Arabic speakers and readers can indeed read Arabic without the dots! The Book Pahlavi script was even much, much worse. Surely a cryptologist would describe it as a "one-way cipher". And yet as recently as the beginning of the 20th century, Zoroastrian high school students in India had to study textbooks to learn how to read ancient Zoroastrian texts in the Book Pahlavi script!

I attach the beginning pages of precisely such a textbook, Lessons in Pahlavi -- Pazend. Part I., "Compiled By Ervad Sheriarji Dadabhai Bharucha, Hon. Fellow of the University of Bombay, and published at the direction of the Trustees of the Parsee Panchayet Funds and Properties, 1908". In the Preface, the author states, "These Lessons are divided into three parts, intended to be learnt by the students of the fourth, fifth and sixth standards respectively of our High Schools." 

These beginning pages then present the characters of the Pahlavi script. Take a look at all of them. Tell me if this is not a "one-way cipher". And yet these secondary school students had to study this textbook and learn how to read it!

If learning to read such a script is possible, then I contend that learning to read with the method I propose for Greek written in the Voynichese script is also possible.

Geoffrey

Another exercise, this one hopefully closer to an actual Voynich MS reading scenario:

You have a text which you know is a pharmacopoeia. You have the following illustration together with Voynichese text on a page:

[attachment=2851]

And you have the following Voynichese text with it:

[  pchor  ros  ochepshod  rosy  oaiin  ochepsho  chaiin  olaly  aiin  s  chkas  ]
[  shol  odcthardy  shaiin  shol  dsheol  tory  chys  s  shod  cthalod  aldary  ]
[  otol  olshy  rchcthol  cheocthy  tar  chaiin  ody  chaiin  tos  shy  olkory  ]
[  shod  char  chod  odar  ody  shaiin  ckhaiin  olky  she  sh2aiiin  toraiin  ]
[  kaiin  tshol  tar  dkeey  shaiin  aiin  s  chkas  dar  chaiin  s  dary  kaiin  olkor  aiin  ]
[  shod  char  chod  chedar  tedshaiin  ]

"[sh2]" is the Voynichese character [sh] with a closed loop on top. All other [sh] characters in this passage of text have an open loop on top.

To help readers get started with the process of reading and understanding this Greek text, I will do the first step of converting the text into the "10-letter Greek alphabet" here:

" pior  rot  oiiptou  rot's  oan  oiipto  `ian  onan's  an  t  `itat  
  ton  opikarus  tan  ton  ption  kor's  `ist  t  tou  kianou  anpar's  
  okon  ont's  riikon  `iioik's  kar  ian  op's  ian  kot  t's  ontor's  
  tou  `iar  iou  opar  ous  tan  tian  ont's  ti  nain  koran  
  tan  kton  kar  ptiis  tan  an  t  `itat  par  ian  t  par's  tan  ontor  an  
  tou  `iar  iou  `iipar  kiptan "

And here is this same text in the 10-letter Greek alphabet in Greek letters:

" πιορ  ροτ  οιιπτου  ροτ'ς  οαν  οιιπτο  `ιαν  οναν'ς  αν  τ  `ιτατ
  τον  οπικαρυς  ταν  τον  πτιον  κορ'ς  `ιστ  τ  του  κιανου  ανπαρ'ς
  οκον  οντ'ς  ριικον  `ιιοικ'ς  καρ  ιαν  οπ'ς  ιαν  κοτ  τ'ς  οντορ'ς
  του  `ιαρ  ιου  οπαρ  ους  ταν  τιαν  οντ'ς  τι  ναιν  κοραν
  ταν  κτον  καρ  πτιις  ταν  αν  τ  `ιτατ  παρ  ιαν  τ  παρ'ς  ταν  οντορ  αν
  του  `ιαρ  ιου  `ιιπαρ  κιπταν "

Note 1: As a reminder that Voynichese final [-y] may often represent a "vowel+s" word ending, I have marked each such case as <'s> or <'ς> in the 10-letter text above.

Note 2: As a reminder that Voynichese initial [ch] may represent either a Greek front vowel or the "rough breathing" /h/ sound, I have marked [ch] in Greek word-initial position as <`i> or <`ι> in the 10-letter text above.

Note 3: This text of 64 Voynichese words represents 52 Greek words.

Note 4: This text is in a Koine Greek dialect, which is more than a millennium earlier and thus much more archaic than the late medieval Greek dialect that the author of the Voynich MS would have spoken and written in. Thus, not every form in this text will precisely match the forms that we would expect to find in the Voynich MS itself. Nevertheless, I think that this text still looks very much like the actual Voynichese of the Voynich MS text.

Note 5: Recall that one of the most important steps in reading the 10-letter Greek text as actual Greek, is to recognize and identify where the letters rho and nu should actually be lambda and mu. In practice I believe that the identification of the lambda's is usually the most critical step, and I find that to be the case in this text as well. It is a universal linguistic tendency for the sounds /l/ and /m/ to be relatively more common in word-initial and syllable-initial position, while /r/ and /n/ tend to be much more common in word-final and syllable-final position. This tendency definitely holds true in Greek, probably to a greater degree than in most other languages, so knowing and using this tendency to help identify lambda's and mu's can potentially be of great assistance to the reader.

Note 6: Remember how "gyro" is pronounced in Greek.

Note 7: The Greek consonant cluster "k+liquid" is written with an unpronounced Voynichese vowel character between the two consonant characters in Voynichese (akin to the way in which Linear B represented such Greek consonant clusters).

Note 8: There is a well-known phenomenon in Greek grammar called "movable nu", by which an optional letter nu is added to the ends of some third person verb forms and dative plural noun forms. In this passage the Voynichese text includes such a movable nu in a few third person verb forms.

Note 9: About a dozen Greek word-final vowels are still deleted in this Voynichese text, but I do not believe this will be an absolute barrier to comprehension. (It is possible that additional "movable nu's" could retain some of these word-final vowels in Voynichese, but I want to proceed sparingly with the use of such movable nu's until I have a more complete understanding of how and to what extent Voynichese may or may not reflect this phenomenon.)

Geoffrey

Quote:Geoffrey: You have a text which you know is a pharmacopoeia. You have the following illustration together with Voynichese text on a page:...

We hope that's what is going on (that there is some correspondence between the text and the images), but it is not yet proven. It could be a journal written within an unfinished herbal manuscript. It could be trade secrets. It could be notes from a spy mission. It could be gibberish.

(14-04-2019, 02:47 AM)-JKP- Wrote: You are not allowed to view links. Register or Login to view.

Quote:Geoffrey: You have a text which you know is a pharmacopoeia. You have the following illustration together with Voynichese text on a page:...

We hope that's what is going on (that there is some correspondence between the text and the images), but it is not yet proven. It could be a journal written within an unfinished herbal manuscript. It could be trade secrets. It could be notes from a spy mission. It could be gibberish.

Of course I agree. In the Voynich MS itself, the connection between the text and the images is not yet proven until a deciphering of the text is accepted by scholars. In my post I just meant that for the purpose of my exercise, we assume that it is a pharmacopoeia, with text related to the illustration. The question at issue is whether my method of encrypting Greek as Voynichese produces such an ambiguous "one-way cipher" that the meaning of the text is irrecoverable to the reader, or whether a reader could still possibly be able to read and understand the Greek text if he or she knows the method. The point of my post is to demonstrate that readers can still be able to read and understand the Greek text if they know the method.

Geoffrey

(14-04-2019, 01:05 AM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view.Another exercise, this one hopefully closer to an actual Voynich MS reading scenario:
You have a text which you know is a pharmacopoeia. You have the following illustration together with Voynichese text on a page
[see previous post for illustration]
And you have the following Voynichese text with it:

[  pchor  ros  ochepshod  rosy  oaiin  ochepsho  chaiin  olaly  aiin  s  chkas  ]
[  shol  odcthardy  shaiin  shol  dsheol  tory  chys  s  shod  cthalod  aldary  ]
[  otol  olshy  rchcthol  cheocthy  tar  chaiin  ody  chaiin  tos  shy  olkory  ]
[  shod  char  chod  odar  ody  shaiin  ckhaiin  olky  she  sh2aiiin  toraiin  ]
[  kaiin  tshol  tar  dkeey  shaiin  aiin  s  chkas  dar  chaiin  s  dary  kaiin  olkor  aiin  ]
[  shod  char  chod  chedar  tedshaiin  ]

"[sh2]" is the Voynichese character [sh] with a closed loop on top. All other [sh] characters in this passage of text have an open loop on top.
To help readers get started with the process of reading and understanding this Greek text, I will do the first step of converting the text into the "10-letter Greek alphabet" here:

" pior  rot  oiiptou  rot's  oan  oiipto  `ian  onan's  an  t  `itat  
  ton  opikarus  tan  ton  ption  kor's  `ist  t  tou  kianou  anpar's  
  okon  ont's  riikon  `iioik's  kar  ian  op's  ian  kot  t's  ontor's  
  tou  `iar  iou  opar  ous  tan  tian  ont's  ti  nain  koran  
  tan  kton  kar  ptiis  tan  an  t  `itat  par  ian  t  par's  tan  ontor  an  
  tou  `iar  iou  `iipar  kiptan "

And here is this same text in the 10-letter Greek alphabet in Greek letters:

" πιορ  ροτ  οιιπτου  ροτ'ς  οαν  οιιπτο  `ιαν  οναν'ς  αν  τ  `ιτατ
  τον  οπικαρυς  ταν  τον  πτιον  κορ'ς  `ιστ  τ  του  κιανου  ανπαρ'ς
  οκον  οντ'ς  ριικον  `ιιοικ'ς  καρ  ιαν  οπ'ς  ιαν  κοτ  τ'ς  οντορ'ς
  του  `ιαρ  ιου  οπαρ  ους  ταν  τιαν  οντ'ς  τι  ναιν  κοραν
  ταν  κτον  καρ  πτιις  ταν  αν  τ  `ιτατ  παρ  ιαν  τ  παρ'ς  ταν  οντορ  αν
  του  `ιαρ  ιου  `ιιπαρ  κιπταν "

Note 1: As a reminder that Voynichese final [-y] may often represent a "vowel+s" word ending, I have marked each such case as <'s> or <'ς> in the 10-letter text above.
Note 2: As a reminder that Voynichese initial [ch] may represent either a Greek front vowel or the "rough breathing" /h/ sound, I have marked [ch] in Greek word-initial position as <`i> or <`ι> in the 10-letter text above.
Note 3: This text of 64 Voynichese words represents 52 Greek words.
Note 4: This text is in a Koine Greek dialect, which is more than a millennium earlier and thus much more archaic than the late medieval Greek dialect that the author of the Voynich MS would have spoken and written in. Thus, not every form in this text will precisely match the forms that we would expect to find in the Voynich MS itself. Nevertheless, I think that this text still looks very much like the actual Voynichese of the Voynich MS text.
Note 5: Recall that one of the most important steps in reading the 10-letter Greek text as actual Greek, is to recognize and identify where the letters rho and nu should actually be lambda and mu. In practice I believe that the identification of the lambda's is usually the most critical step, and I find that to be the case in this text as well. It is a universal linguistic tendency for the sounds /l/ and /m/ to be relatively more common in word-initial and syllable-initial position, while /r/ and /n/ tend to be much more common in word-final and syllable-final position. This tendency definitely holds true in Greek, probably to a greater degree than in most other languages, so knowing and using this tendency to help identify lambda's and mu's can potentially be of great assistance to the reader.
Note 6: Remember how "gyro" is pronounced in Greek.
Note 7: The Greek consonant cluster "k+liquid" is written with an unpronounced Voynichese vowel character between the two consonant characters in Voynichese (akin to the way in which Linear B represented such Greek consonant clusters).
Note 8: There is a well-known phenomenon in Greek grammar called "movable nu", by which an optional letter nu is added to the ends of some third person verb forms and dative plural noun forms. In this passage the Voynichese text includes such a movable nu in a few third person verb forms.
Note 9: About a dozen Greek word-final vowels are still deleted in this Voynichese text, but I do not believe this will be an absolute barrier to comprehension. (It is possible that additional "movable nu's" could retain some of these word-final vowels in Voynichese, but I want to proceed sparingly with the use of such movable nu's until I have a more complete understanding of how and to what extent Voynichese may or may not reflect this phenomenon.)

I will offer a guide to the first line to help readers get started. I think readers would be able to do this process by themselves on their own, with some practice and experience.

You have the illustration, and in the first line in the 10-letter Greek text you see the repeated root "ροτ" and "ροτ'ς" ("rot" and "rot's"). Knowing that rho/lambda substitution is one of the first things you should check in this reading process, I think the meaning of "ροτ'ς" will be very readily apparent to the reader, knowing what the text is about and seeing the illustration.

You also have the repeated root "οιιπτου" and "οιιπτο" ("oiiptou" and "oiipto"). Having recognized "ροτ'ς", I think the reader can also recognize "οιιπτου" rather readily. (The "gyro" rule is helpful here.)

Point: with a modicum of common sense about the context on the reader's part, these words are not actually ambiguous at all. 

Now once the reader has gotten started with the plant name words at the beginning of the text, the Greek function words around them will begin to fall into place. Then, toward the end of the line, the experienced reader can see that you clearly have a prepositional phrase at the end of the line: "[prep.] [article] [noun]".

Using this context clue, the reader can figure out that a verb or participle form with a verbal meaning is very likely to occur in between the plant name/location and the prepositional phrase. With this expectation, and with the aid of nu/mu substitution and the "gyro" rule again, the experienced reader can figure out the verb/participle word in this place in the text.

Thus, using context in the proper way, the entire first line will not be ambiguous at all to a reader with some practice and experience with such a method and writing system.

And once one has begun reading and comprehending entire lines of text, it will become increasingly possible for the experienced reader to use context to guess what the following words and phrases are likely to be anyway. Having such context, the 10-letter Greek alphabet form of the words will be enough for the reader to figure out what the rest of the text says. 

I have in mind of course a late medieval Greek speaker and reader for whom reading and using such a pharmacopoeia was a regular part of their daily life experience and practice. It is more difficult for us in modern times, even for a native speaker, since we no longer need to use such pharmacopoeias on a regular basis. But if you lived in the 15th century, and the life and health of your family members depended on your ability to read and understand the pharmacopoeia, I bet you could learn to become surprisingly proficient in reading and understanding the meaning of such a text even in a 10-letter alphabet form.

Geoffrey

(continued)

I’ll go on from where I dropped last time. The next concern I have is:

2. Inconsistency in writing the same word into Voynich letters.

I don’t know if you have answered this question somewhere else, so you can just give me the link to the answer if you have done.

The question arises in the first line of You are not allowed to view links. Register or Login to view., where the same Greek word [font=Noto Serif, Cambria, serif]είπαν (eípan) appear twice:[/font]

Quote:first three lines of You are not allowed to view links. Register or Login to view. 1 in the Voynich ms text:

[t]eeodaiin  shey  epairody  osaiin  yteeoey  shey  epaiin  oaiin
daiir  okeody  qoekeeg  sar  oeteody  oteey  keey  key  keeodal
ycheo  s oeeg  cheos aiin  okesoe  aram  shees  dalaiin  dam

my Judaeo-Greek interpretation of this text:

[]ei[A]pan  tis  ipeirous  otan  skiiAis  tis  ,  epan  oAn
par'  Atous  &Atees  tAr(a)  oikous  o(u)k-eis(i)  tees  ,  tis  t-ei[A]pAs
shio  t-Aees  heAt-an  Atitoi  AlAs  ,  deit  vasAn  ,  fAs

Now "normalizing" this Judaeo-Greek text into a more standard Greek form:

eipan  tis  ipeirous  otan  skiais  tis  ,  eipan  oun
para  autous  &  autes  tora  oikous  ouk  eisi  tes  ,  tis  t'-eipes
sou  tes  etan  auttoi  aules  ,  deite [te] vasein  ,  phes

For me, it would be natural to expect the same word to appear in the same form, because when we read, we do not read each word letter-by-letter, but You are not allowed to view links. Register or Login to view.. The two words that you interpreted as [font=Noto Serif, Cambria, serif]είπαν are written in Voynichese as eeodaiin and epaiin, respectively.

They look quite different to me, especially the fact that the same Greek letter π (p) are written in Voynichese as d and p, respectively. It is hard for me to believe that they are the same word. I’d like to see an explanation here.

(to be continued…)[/font]

Thank you, Zhe. This is a serious argument against Geoffrey's theory, since one of his main lines of defense is that native speakers would get used to the system. As you say, this would be difficult with varying word images. If anyone should be able to read this, they should be able to develop a consistent mental word list.

The issue of 'eipan' was addressed here: You are not allowed to view links. Register or Login to view.

It is the immediate consequence of working 'backwards' i.e. concentrating on the Voynich to plaintext conversion rather than the other way round.

(continued)

3. [d] as a substitute for [p] and [f]

From You are not allowed to view links. Register or Login to view., it seems that you propose [font=Eva]d as a possible (but not purely) substitute for p and f, because of:
[/font]

Quote:

• d may be representing /u~v/, while
  
• p and f may be representing /b~f~p/, and
  
• Betacism may caused the writer to write /v/ for /b/, and from where it was extended to writing /v/ for /f~p/
  

The last line does not seem likely to happen to me, but I cannot say it’s impossible. However, Emma then posted these statistical facts:

Quote:

• [d] is common on the first line of a paragraph, which would need to be explained.
  
• [p, f] are shorter when written away from the first line of paragraphs and [k, t] are taller when written on the first line. The scribe was capable of adjusting the glyph height.
  
• [d, l, r, s] all lack significant number of [e] following them.
  
• [p, f] take relatively good numbers of [ch, sh] following them, which [d] does not, especially away from the start of the line.
  

So, it seems that the [font=Eva]d cannot replace p and f freely, as they have distinct stats.

Again, the question arises in the first line of You are not allowed to view links. Register or Login to view., where the same Greek word είπαν (eípan) appear twice, but the same Greek letter π (p) are written in Voynichese as d and p, respectively.

Since the stats are different for d, p and f, it would be great if we could have a general rule about when to use d. I guess the final answer must be based on very large amount of texts, so that the rule could be general enough to cover most cases. Therefore, I’m not expecting to see an answer soon, but when you have finished translating the whole manuscript, there should have already been an answer to this question.

(to be continued…)[/font]

(18-04-2019, 07:46 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.The issue of 'eipan' was addressed here: You are not allowed to view links. Register or Login to view.

It is the immediate consequence of working 'backwards' i.e. concentrating on the Voynich to plaintext conversion rather than the other way round.

Thanks, René. It seems that Geoffrey’s explanation about different [font=Noto Serif, Cambria, serif]είπαν (eípan) is like this:[/font]

(09-04-2019, 10:01 PM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view.The most significant difficulty in explaining the vowel representations in these forms is the presence of Voynich [o] between [ee] and [d] in the first and third cases [(t)eeodaiin] and [keeodal]. The only thing I can say about this at the present stage is that I note the well-known (to Voynich researchers) distinction between "Currier A dialect", which almost never writes the combination [ed], and "Currier B dialect", which very frequently writes the combination [ed]. But Currier A dialect does use the combination [eod]. So these forms here appear to reflect the "Currier A dialect" style, avoiding the combination [ed].

Geoffrey

I have to say that this is not persuasive. Before explaining why there is an [o] between [ee] and [d], I would like to see why [d] is used here instead of [p] first. As I posted above, the idea that [d] being substitute for [p] and [f] is not so firm by itself.

Moreover, in my opinion, why combination [ed] has to be avoided is another question that has to be answered for the whole theory to work.