+- 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: Experimental replica of VMS properties with a given corpus (/thread-2737.html)
RE: Experimental replica of VMS properties with a given corpus - Aldis Mengelsons - 13-04-2019
(13-04-2019, 01:48 PM)-JKP- Wrote: You are not allowed to view links. Register or Login to view.(13-04-2019, 12:24 PM)Aldis Mengelsons Wrote: You are not allowed to view links. Register or Login to view.[ From which page did you took that text??? translating its kinda interesting
Were you talking about the medieval sample, Aldis? It's folio 101 (original foliation) BAV Pal. Lat. 25.
Ok! Thanks a million man!
Because the text was very interesting,especially middle part
"you disbelieve to your teacher which sharing his knowledge for free.Now you holding that and its feel sweet as i see"
Thaks!Iwill do 101 in very next day,maby something exiting will "pop up".)))
RE: Experimental replica of VMS properties with a given corpus - bi3mw - 14-04-2019
When checking the word length in the VMS I received a very strange result. That does not look like one natural language to me, but as something "constructed". The result is the following:
![[Image: vms_word_length.png]](http://wwwhomes.uni-bielefeld.de/mwille2/VMS/vms_word_length.png)
Does anyone know any natural language that causes a similar result ? I always get much flatter, more homogeneous curves.
You are not allowed to view links. Register or Login to view.
RE: Experimental replica of VMS properties with a given corpus - geoffreycaveney - 14-04-2019
(14-04-2019, 11:53 AM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.When checking the word length in the VMS I received a very strange result. That does not look like one natural language to me, but as something "constructed". The result is the following:
Does anyone know any natural language that causes a similar result ? I always get much flatter, more homogeneous curves.
It could still be natural language, but the author could have placed Voynichese word breaks in different places than the natural language's word breaks. Prefixes and/or suffixes could be written as separate words, while short prepositions, conjunctions, and particles could be written as part of the following word. Also, just as in scripts like Arabic, Persian, and Syriac/Aramaic, certain letters are never "connected" to the following letter, the author could have automatically placed a Voynichese word break after certain characters. But despite all of these irregularities in the word breaks, it could still represent natural language.
Geoffrey
RE: Experimental replica of VMS properties with a given corpus - farmerjohn - 14-04-2019
(14-04-2019, 11:53 AM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.When checking the word length in the VMS I received a very strange result. That does not look like one natural language to me, but as something "constructed". The result is the following:
Does anyone know any natural language that causes a similar result ? I always get much flatter, more homogeneous curves.
Stolfi has great set of articles:
in particular You are not allowed to view links. Register or Login to view. is about word length distribution and You are not allowed to view links. Register or Login to view. is about candidate languages
RE: Experimental replica of VMS properties with a given corpus - bi3mw - 14-04-2019
Well, I have no idea about Chinese, Vietnamese or Tibetan manuscripts, but the illustrations in the VMS seem more European to me. (e. g. swallowtail merlons in Asia ?).
I am very skeptical of an Asian origin. A whatever created construction seems more likely to me.
RE: Experimental replica of VMS properties with a given corpus - Koen G - 14-04-2019
(14-04-2019, 01:37 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.I am very skeptical of an Asian origin. A whatever created construction seems more likely to me.
I've grown more and more convinced that everything about the VM is a construction rather than a natural cultural product. So I agree with the bit "a whatever created construction". But the fact that it is artificial makes Asian languages more likely instead of less.
We're dealing with an encrypted text (if it contains meaning), so a scenario where a secretive or oppressed fringe group camouflages a text in their language in a Latinesque script next to Latinesque drawings is a possibility.
I'm not saying that an Asian language is the most likely option, but if the VM were a spontaneous cultural utterance then Asian languages would be much less likely than in the case of a construct.
RE: Experimental replica of VMS properties with a given corpus - bi3mw - 14-04-2019
It is also conceivable that the words are not separated as you would expect, but like this:
thiss ente ncei sjust anexa mpl efori rregu larc utting
Before that, of course, another step must be taken. Or the text actually consists of extreme abbreviations that are sufficient as "encryption".
RE: Experimental replica of VMS properties with a given corpus - MarcoP - 14-04-2019
(14-04-2019, 03:56 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.It is also conceivable that the words are not separated as you would expect, but like this:
thiss ente ncei sjust anexa mpl efori rregu larc utting
Before that, of course, another step must be taken. Or the text actually consists of extreme abbreviations that are sufficient as "encryption".
This kind of arbitrary spacing is one of the many tweakings in Geoffrey's theory.
But You are not allowed to view links. Register or Login to view. are many more than the additional ones you get by ignoring spaces (47% vs 8%). And what about reduplication? Are repeating words "nulls" as Mark thinks?
Hammering Voynichese into Latin (or Greek) soon leads to rejecting larger and larger parts of the evidence.
I prefer to keep an open mind on the possible language, trying to consider all evidence as meaningful.
PS: Whoops, we are off topic again
RE: Experimental replica of VMS properties with a given corpus - geoffreycaveney - 14-04-2019
(14-04-2019, 04:10 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.(14-04-2019, 03:56 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.It is also conceivable that the words are not separated as you would expect, but like this:
thiss ente ncei sjust anexa mpl efori rregu larc utting
Before that, of course, another step must be taken. Or the text actually consists of extreme abbreviations that are sufficient as "encryption".
This kind of arbitrary spacing is one of the many tweakings in Geoffrey's theory.
But You are not allowed to view links. Register or Login to view. are many more than the additional ones you get by ignoring spaces (47% vs 8%). And what about reduplication? Are repeating words "nulls" as Mark thinks?
Hammering Voynichese into Latin (or Greek) soon leads to rejecting larger and larger parts of the evidence.
I prefer to keep an open mind on the possible language, trying to consider all evidence as meaningful.
PS: Whoops, we are off topic again
A brief off-topic criticism of my theory deserves a brief off-topic response:
Writing some suffixes and syllables as separate words, combined with the ambiguity of letter values of characters, actually offers a quite natural explanation of much of the reduplication of words in the text:
If many syllables are written separately, and if sequences such as /ten/, /dyn/, and /thain/ can be written identically as homographs, then it is much more likely that these syllables will occur consecutively in a text, creating the appearance of reduplication of Voynichese words.
Voila, another unusual feature of the text explained.
Yes, the ambiguity creates challenges for the reader to comprehend the text, but we are discussing that issue with a series of exercises in the other thread.
Geoffrey
RE: Experimental replica of VMS properties with a given corpus - -JKP- - 14-04-2019
I don't think reduplicated words are nulls. Not only does the text not appear to work that way, but it would take some of the blocks out of the VMS's already limited glyph-set.
In fact, in years of looking at the text (and trying to manipulate it in a thousand ways), there's only one character that seems to me to behave like a null and even that one I'm not sure of (maybe it's a letter, marker, or modifer). I have a couple of non-null ideas (one in particular) that might explain the behavior of that particular glyph.
I do think the spaces are contrived. I think VMS tokens are probably syllables in some sense, not words. I've suspected it for a long time, have never quite been able to confirm it, many have disagreed with me, but some work I've been doing the last couple of years seems to point even more in that direction. I'll keep pounding away at it until I have an answer.
.
Reduplication can be explained in many ways and over the years I've been collecting manuscript samples with a large amount of repetition and reduplication, and there's more of it than I suspected. When I have time, I'll post the examples. Even if the VMS is not reduplication in the same sense as the examples (which are of many different kinds), reduplication can be simulated simply by altering spaces and grouping glyphs as multiglyphs.
As for multiglyphs, I've written about this on blogs and have posted charts and I've been investigating it for two reasons: Studying possible multiglyphs will either eventually reveal if they exist and what they are or, if they are sparse or nonexistent, they will hopefully reveal enlightening patterns (some of which hopefully are linguistic in nature).
In fact, it was my study of multiglyph patterns (without the assumption that they are multiglyphs, but just looking at them as patterns) that led me to the language I used to create the almost-substitution sample I posted in René's thread (creating Voynich-like text from natural language).
.
If there are multiglyphs, it would nullify some of the results of computational attacks that assume one-to-one correspondence between VMS glyphs and alphabetic letters, but the possibility of multiglyphs has to be considered, and if the studies have to be redone, they have to be redone. I don't get the feeling the VMS is highly verbose, however. The patterns don't seem to lean that way (more study is needed but the repeating patterns within them that don't seem to behave as highly verbose, as far as I can see... so far).