09-10-2024, 06:37 AM
I just finished watching the video and find its presentation very clear, one of the trademarks of Koen's videos.
The problem presented here is real, and it explains one of the reasons for the low entropy in non-mathematical terms.
This does not mean, in my humble opinion, that an explanation of this feature has to be very far-fetched or even complicated.
We need to remember that the correspondence between sounds (phonemes) and the letters of the alphabet (graphemes), can have quite a loose relationship, which depends heavily on the language.
Spoken Dutch and written Dutch are quite different, but it cannot beat English in that respect.
There are more sounds than letters. Some of these (e.g. diphthongs) may not be considered single sounds - I suppose that linguists have a definition for that.
However, some languages use single symbols (letters) for diphthongs and others use two, three or more characters for a single phoneme. In fact, English does all of this.
Furthermore, the same sound may be represented by different groups of characters, and the same character may produce several different sounds. Again, English does all of that.
We are perhaps fortunate that the MS is not written in modern English....
Czech has a longer alphabet than English, and has a reasonably close mapping between characters and sounds.
The sound represented by English 'sh' (two characters, one phoneme, except in some words like 'mishap'), is represented in German by three (sch) and in Czech by one (š).
Thus, a small alphabet can represent a larger sound inventory by using combinations.
If that is really what is happening in the MS, then words in the Voynich MS, which are already short, become even shorter, which is another problem.
This suggestion can possibly solve one problem, but several others remain.
In particular, the point of the video, that simple substitution (or something largely based on it) cannot work, remains completely valid.
Solvers 'in spe' who base their solution on simple substitution cannot fail to realise that it does not really work - it generates only some valid words and never valid phrases. They tend to go to one of the following modifications (or both):
- suggest that the plain text is in a mixture of languages (Levitov, Cheshire)
- suggest that some Voynich symbols can represent several different plain text characters (too many to name)
Both lead to more words, but still no valid phrases.
Both also could conceivably seem sensible when only thinking about the back translation from Voynichese to plain text, but barely make sense when considering the act of creating the MS text.
Who would have spoken in an arbitrary mixture of several languages? Certainly not more than two, and only in a consistent manner (e.g. vernacular mixed with Latin).
Who would have encoded both an 'r' and a 'd' with the same cipher symbol?
It is clear that that is just a device by the decoder in order to match more words...
The problem presented here is real, and it explains one of the reasons for the low entropy in non-mathematical terms.
This does not mean, in my humble opinion, that an explanation of this feature has to be very far-fetched or even complicated.
We need to remember that the correspondence between sounds (phonemes) and the letters of the alphabet (graphemes), can have quite a loose relationship, which depends heavily on the language.
Spoken Dutch and written Dutch are quite different, but it cannot beat English in that respect.
There are more sounds than letters. Some of these (e.g. diphthongs) may not be considered single sounds - I suppose that linguists have a definition for that.
However, some languages use single symbols (letters) for diphthongs and others use two, three or more characters for a single phoneme. In fact, English does all of this.
Furthermore, the same sound may be represented by different groups of characters, and the same character may produce several different sounds. Again, English does all of that.
We are perhaps fortunate that the MS is not written in modern English....
Czech has a longer alphabet than English, and has a reasonably close mapping between characters and sounds.
The sound represented by English 'sh' (two characters, one phoneme, except in some words like 'mishap'), is represented in German by three (sch) and in Czech by one (š).
Thus, a small alphabet can represent a larger sound inventory by using combinations.
If that is really what is happening in the MS, then words in the Voynich MS, which are already short, become even shorter, which is another problem.
This suggestion can possibly solve one problem, but several others remain.
In particular, the point of the video, that simple substitution (or something largely based on it) cannot work, remains completely valid.
Solvers 'in spe' who base their solution on simple substitution cannot fail to realise that it does not really work - it generates only some valid words and never valid phrases. They tend to go to one of the following modifications (or both):
- suggest that the plain text is in a mixture of languages (Levitov, Cheshire)
- suggest that some Voynich symbols can represent several different plain text characters (too many to name)
Both lead to more words, but still no valid phrases.
Both also could conceivably seem sensible when only thinking about the back translation from Voynichese to plain text, but barely make sense when considering the act of creating the MS text.
Who would have spoken in an arbitrary mixture of several languages? Certainly not more than two, and only in a consistent manner (e.g. vernacular mixed with Latin).
Who would have encoded both an 'r' and a 'd' with the same cipher symbol?
It is clear that that is just a device by the decoder in order to match more words...
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)