+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Analysis of the text (https://www.voynich.ninja/forum-41.html)
+--- Thread: [split] Verbose cipher? (/thread-3356.html)
RE: [split] Verbose cipher? - geoffreycaveney - 18-09-2020
Here is an example of what my You are not allowed to view links. Register or Login to view. looks like for a short passage of actual Voynich ms text. The following text comprises the four lines at the top of ms page f67r1, transcribed to represent a modified form of Koen's verbose cipher as I interpreted and adjusted it:
[p]égo r'ey ePálgy hro ypéhey r'ey ePo ho
kál degy neté[y] ral hepegy béy téy tey téga
yjeh r hé[y] jehr o derhe al á r'ér kao ká
jego aetéy ral ál rhal jéy kál p'ey
I find the "floating" EVA [o]'s (not part of one of the [o+glyph] bigrams) interesting when represented as <h>'s, as the VCI transcription does.
I also must note that it is very likely that this transcription system, like others, is oversimplifying the characters transcribed as EVA "[sh]", and here transcribed as <r'>. I strongly suspect this actually represents multiple letters, depending on whether the loop on top is open, closed, or perhaps a third form.
I also note that EVA "[d]" may actually represent multiple forms and letters as well, so its transcription here as <k> (and of EVA [od] as <g>), may also be an oversimplification.
The VCI transcription system currently lacks such values as <c>, <č>, and <x> (the voiceless velar fricative phoneme), so they may well be among the other actual values of different forms of EVA [sh] and EVA [d].
Geoffrey
RE: [split] Verbose cipher? - Koen G - 18-09-2020
That's really interesting, Geoffrey. I'm surprised at how "normal" you made it look
RE: [split] Verbose cipher? - MichelleL11 - 18-09-2020
(18-09-2020, 05:40 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.That's really interesting, Geoffrey. I'm surprised at how "normal" you made it look
My thoughts as well. Thank you for sharing your efforts with the board.
RE: [split] Verbose cipher? - geoffreycaveney - 18-09-2020
I'm glad it seems interesting to you, Koen and Michelle.
Someone, I believe it may have been Marco, raised the point that a verbose cipher could make the labels in the Voynich ms even more difficult to interpret, as each label would carry even less content in terms of its actual number of distinct meaningful units. Out of curiosity I tested the VCI transcription on the 12 labels around the outside of what may be a Zodiac circle on folio page f67v1. I consider the results potentially rather encouraging in terms of the "readability" of the labels as well as their distinction from each other (as opposed to the apparently excessive repetition seen in their EVA transcription readings). To wit:
(Starting with EVA [okal ary] near 12 o'clock position and proceeding clockwise)
da aly
kl itza
dzaky
kirytiy
ytegal
keihets
d'hrá
thialky
higale
zr oy
káás
kálta
I continue to be pleased at the impact of treating independent EVA [o]'s as VCI <h>'s.
Geoffrey
RE: [split] Verbose cipher? - Koen G - 18-09-2020
(18-09-2020, 07:09 PM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view.as each label would carry even less content in terms of its actual number of distinct meaningful units.
I'm not sure if that is the best way to put it. At least within the mindset of the verbose cipher, the two glyphs in something like [in] are hardly worth considering distinct meaningful units. It's the same soup, but less watered down
Or to put it the other way, the supposedly low information content of Voynichese is something to worry about either way.What Marco noticed was that words simply get really short on average. But I agree that the label examples you show look promising! I still consider this completely an experiment, not a theory, and I'm very happy people like you have a go at it as well.
RE: [split] Verbose cipher? - Anton - 18-09-2020
Normalizing entropy is an interesting and worthwile task in itself, and I was enthusiastic about it myself a few years ago, when there were first discussions on entropy in the Forum. But... somehow it's easy to over-focus upon it, and it's as easy to lose grasp of the "big picture" - just as a substitution-cipher-solver would lose it (the difference is that the substitution-cipher-solver is most probably yet unaware of various properties of Voynichese)
IMO, one should always bear in mind the "big picture", -which is probably not altogether accomplished yet, - but its components are:
1) entropy
2) positional properties of characters within vords
3) positional properties of characters within lines and paragraphs
4) gallows coverage
5) vord reduplication (strict and quasi-)
6) multipass
7) graphical decomposition of characters
8) <fill in your own variant here>
RE: [split] Verbose cipher? - RenegadeHealer - 19-09-2020
Nice work, Geoffrey. I'm intrigued and entertained. Whether or not it pans out, what you've built is a consistent, logical, testable model for phonetically mapping Voynichese characters, built on solid facts about the properties of the text. Also, from my perspective as a self-taught amateur at linguistics, I'm seeing a solid understanding of the basics of phonology. It's at least a good start; congratulations on constructing a model worth testing (and able to be tested!) at all.
This is probably a matter of taste and personality, but seeing people present fleshed-out ideas about the VMs, and then receive critical feedback, does not, per se, make me cringe. I only cringe when someone receives feedback on their ideas less than gracefully. Like in judo, there is much to be learned by taking a swing at the VMs and missing, understanding why you missed, and observing others do the same. I don't think anybody should be shamed for trying, as long as they receive all feedback graciously and seriously.
A couple of questions, Geoffrey:
- Where do you currently stand on the idea of EVA [f] and/or [p] being top-line variations of another glyph? In your model, you equate [f] with [k] and [p] with [t]. Last year, you and I had been toying with the idea of [f] and [p] actually being top-line variations of [d]. What made you change your mind?
- I'm wondering if EVA [dch] also belongs in your chart, as /kʲ/
- What are your thoughts on the controversial but intriguing possibility of the equivalence of EVA [a] and [y]?
- Suppose for a moment that [a] with no glyphs attached to the right of it — to the extent that this glyph can even be said to have an independent existence — is the same as [y]. Koen and Marco have presented evidence for each in the series [a*] being a separate letter, and your model expands this series into a set of vowel sounds. Wouldn't it make sense, then, that isolated [a] and [y] have the same pronunciation, and both of them represent the most neutral or empty vowel possible? /a/ is definitely a possibility, but I wonder if some type of schwa or barely-pronounced vowel sound could also be it. I'm thinking of the way the Cyrillic alphabet's various yers have lost their status as true vowels, or the way Mandarin Chinese Pinyin Romanization uses "i" after a sibilant to indicate the near lack of a vowel. Or, for a more familiar example to many here, the way Germans pronounce the letter "e" at the end of a word. I also wonder if unattached [a] or [y] could be a wildcard vowel sign, which can be used to write a variety of different vowels, perhaps unaccented ones.
- If the series [o*] represents a series of voiced consonants, might there be a corresponding series of [y/a*], used to explicitly specify an unvoiced consonant? [yk], [yt], [yd], [ar], [as], [al]
- If [or] represents /l/, might [ar] then represent /ɫ/? This is a phonemic distinction in some of the languages under consideration, after all.
- I'm not quite following your logic behind [a] = /a/ and [l] = /s/, but then [al] = /a:/. I'm reminded of how in French, syllable coda "s" is not pronounced, but still lengthens the preceding vowel
RE: [split] Verbose cipher? - DONJCH - 19-09-2020
I am surprised that f67v1 is interpreted as a 12 segment zodiac when there are at least 2 more segments that run off the page.
RE: [split] Verbose cipher? - MarcoP - 19-09-2020
(18-09-2020, 10:28 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.Normalizing entropy is an interesting and worthwile task in itself, and I was enthusiastic about it myself a few years ago, when there were first discussions on entropy in the Forum. But... somehow it's easy to over-focus upon it, and it's as easy to lose grasp of the "big picture" - just as a substitution-cipher-solver would lose it (the difference is that the substitution-cipher-solver is most probably yet unaware of various properties of Voynichese)
IMO, one should always bear in mind the "big picture", -which is probably not altogether accomplished yet, - but its components are:
1) entropy
2) positional properties of characters within vords
3) positional properties of characters within lines and paragraphs
4) gallows coverage
5) vord reduplication (strict and quasi-)
6) multipass
7) graphical decomposition of characters
8) <fill in your own variant here>
Thank you, Anton!
I would add "the drift from Currier A to Currier B".
It's depressing to see verbose ciphers treated exactly like all the delusional attempts at simple substitution. The fact that verbose ciphers are trillions of times more complex seems to go unnoticed. Actually, the added complexity risks being welcome to the delusional solver, who only sees more room to play around.
RE: [split] Verbose cipher? - nablator - 19-09-2020
My 2 € cents: tokenization is not the only way to deal with the entropy problem. Personally I rejected tokenization (and therefore verbose ciphers) because Voynichese is too rich and inventive to be harnessed by fixed rules dealing with special properties of some strings of glyphs. Even in an "add new rules whenever you want" scenario, some consistency would have to be maintained and it would be a difficult (and boring) task to maintain and apply consistently a list of evolving rules, especially if several scribes are involved.