additionally, in oct 2016, I've made an You are not allowed to view links.
Register or
Login to view.on the most frequent letters per word and what letter makes a
word unique.
It appears that in every word you can found at least one of these letters from the group [o c a y ]. If you remove these words,
then there are only 170 words remaining, such as [dl].
(03-04-2019, 01:00 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.It appears that in every word you can found at least one of these letters from the group [o c a y ].
That's an interesting way of looking at it. One might say that on the surface this is an indication of the vowel-like status of these glyphs? Remove a,e,i,o,u words from Spanish and there's not much left either. So this is surely one of the reasons why Voynichese feels like a language.
(03-04-2019, 01:18 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.So this is surely one of the reasons why Voynichese feels like a language.
There is considerable quantitative evidence (shifts and jumps in glyph n-gram statistics, autocopy, LAAFU, PAAFU, entropy increase on the third glyph, etc.) that the text is not what it appears (superficially) to be: written in an unknown but not-too-weird natural language. Also, beside vowel-like glyphs that happen to work well as vowels, it looks like the author(s) put far too much in their creation of what makes a text language-like (repeated words, common glyph patterns and common Gothic abbreviations) to make it as obvious as possible that it is language, nothing but language.
If the VMS is inherently deceptive about the nature of the text, either because it is strongly coded or encrypted (an anachronism in the 15th century), or because it represents something else than language, then everything else
prima facie about the VMS (an otherwise unremarkable herbal/astrological/medicinal manuscript) is probably nothing but deception and misdirection too. It would be inconsistent to go through the effort of creating a unbreakable system for the text and then label objects and give precious hints about the subject in illustrations, giving researchers a "low hanging fruit" to attack.
(05-04-2019, 12:18 PM)nablator Wrote: You are not allowed to view links. Register or Login to view. (03-04-2019, 01:18 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.So this is surely one of the reasons why Voynichese feels like a language.
There is considerable quantitative evidence (shifts and jumps in glyph n-gram statistics, autocopy, LAAFU, PAAFU, entropy increase on the third glyph, etc.) that the text is not what it appears (superficially) to be: written in an unknown but not-too-weird natural language. Also, beside vowel-like glyphs that happen to work well as vowels, it looks like the author(s) put far too much in their creation of what makes a text language-like (repeated words, common glyph patterns and common Gothic abbreviations) to make it as obvious as possible that it is language, nothing but language.
If the VMS is inherently deceptive about the nature of the text, either because it is strongly coded or encrypted (an anachronism in the 15th century), or because it represents something else than language, then everything else prima facie about the VMS (an otherwise unremarkable herbal/astrological/medicinal manuscript) is probably nothing but deception and misdirection too. It would be inconsistent to go through the effort of creating a unbreakable system for the text and then label objects and give precious hints about the subject in illustrations, giving researchers a "low hanging fruit" to attack.
To me the most obvious example of a "low hanging fruit" is the illustration of the group of seven stars. The MS author made no attempt to disguise the intended signification of this illustration: it is obvious to everyone on Earth for all of history that it must represent the Pleiades. We recognize it; Aristotle would recognize it; the author clearly must have known everyone would recognize it. And the author put a word label right next to this illustration.
On the other hand, the Zodiac symbol illustrations on their separate pages do
not have obvious word labels right next to each of them. The only obvious labels for them are the ones written in later in Latin letters, apparently in a dialect of French.
If the author wanted to avoid giving any clues, he or she would not have put a word label right next to the illustration of the group of seven stars. On the other hand, if the author wanted to give deliberately false clues with deceptive illustrations, I would think he or she would have done so with the Zodiac symbol illustrations in the same way that he or she did with the Pleiades illustration.
The lack of consistency in the handling of the labeling of the Pleiades vs. that of the Zodiac symbols suggests to me that there was not some systematic deliberate plan intended on the part of the author throughout the MS text, either to avoid providing clues or to deliberately provide false clues as a form of deception.
Geoffrey
(03-04-2019, 01:00 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.additionally, in oct 2016, I've made an You are not allowed to view links. Register or Login to view.on the most frequent letters per word and what letter makes a word unique.
It appears that in every word you can found at least one of these letters from the group [o c a y ]. If you remove these words,
then there are only 170 words remaining, such as [dl].
Davidsch, thank you for this analysis. It is very interesting.
I should make it clear to everyone that unlike many other people on this forum and involved in Voynich research, I am not a computer programmer or a coder at all. My angle to approach the MS text is mainly philological, with a focus on the linguistics, grammar, morphology, phonology, etc., of the possible underlying target languages of the MS text. But my focus is not on "computational" linguistics per se, although I am aware and I recognize that this is an active field of linguistic research as well.
As such, many things that may seem obvious or simple to many of you when reading Davidsch's blog post, may not seem obvious or simple to me. So I am going to ask some questions about it, even though the answers may already be obvious or simple to many of the rest of you.
Am I correct in understanding that your analysis indicates that [o] is the first letter that occurs from among the group [o c a y ] in 16070 words? And so on with the other letters. But there are many other words in which [o] occurs, just not before all of [c a y ]? Thank you very much for clarifying this point if you can.
I am curious about what happens to the statistics when you remove [y] from the combination group of high ranked letters. Those who are following my own pet theory will easily guess why I am asking about removing [y] in particular, but let me focus on the abstract analytical aspect of the question here. I imagine based on your analysis that in this case the best remaining 4 letter combination will be [o c a e ]. Now if you had to add a 5th letter, but you were not allowed to add [y], which would be the 5th letter? Would it be [d]? This is my guess, but I do not want to jump to conclusions.
Thank you for your patience with my questions, and for taking the time to consider them.
Geoffrey