The Voynich Ninja

Many people who have studied the text of the Voynich Manuscript have come to see the character [n] as not different or independent from [i]. Either it they see it as simply a final variant or they see [in], [iin], and [iiin] as single characters. I'm slowly believing that one of these is likely to be true.

But how can we prove that this is the case? The three bits of evidence I would put forward are:

1) the strong link between [i] sequences and [n]: over 90% of [i] are followed by another [i] or [n], and something like 97% of [n] are preceded by [i]; and
2) [i] and [n] have complementary distribution at the end of words: [i] is practically never word final while [n] is practically always word final; and
2) both [i] and [n] greatly prefer following [a] than [o]: the ratio is 20:1 and 25:1 respectively.

What else could we say to strengthen the case, and what would prove it?

I find the second option, where in/iin/iiin are different characters, more plausible because it would likely increase the character entropy, whereas an equivalence between 'i' and 'n' would decrease it even further.

The 'iin' stroke seems to occur occasionally in manuscripts of the time (it's not rare), and can indicate an abbreviation.

The following are some examples from Cappelli.

'iin' as a 'u' in 'anuu', encoding 'annuum':
You are not allowed to view links. Register or Login to view.

'iin' as a 'u' in 'qartu', encoding 'quartum':
You are not allowed to view links. Register or Login to view.

'in' as an 'i' in 'fli', encoding 'falsi':
You are not allowed to view links. Register or Login to view.

'in' an an 'i' in 'sait', encoding 'sanitatis':
You are not allowed to view links. Register or Login to view.

The 'n' stroke also occurs as extension to 'a'. Here it is in 'spear', encoding 'specialiter':
You are not allowed to view links. Register or Login to view.

In 'subaliu', encoding 'substantialium', the 'n' stroke occurs twice:
You are not allowed to view links. Register or Login to view.

The result can be a complex mix where:
'i' takes the form of EVA 'i' or 'in'
'u' takes the form of EVA 'ii' or 'iin'
'iu' takes the form of 'EVA iii' or 'iiin'

And so on.

Job, your point about entropy is very interesting, and something I had not yet considered. I agree that this makes the latter option much more welcome.

From your examples, do you feel that [n, in, iin, iiin] are likely to have substantially different values? Or even differing values depending on the word?

(19-02-2016, 07:16 PM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.From your examples, do you feel that [n, in, iin, iiin] are likely to have substantially different values? Or even differing values depending on the word?

I think you can make a case for it, but i haven't formed an opinion.

For example, in You are not allowed to view links. Register or Login to view. (nicely covered by Brian Cham You are not allowed to view links. Register or Login to view.), 'Agrimonia' is written in this way:
[attachment=118]

Characters 'm', 'n' and 'i' are composed of the same stroke. The same could be true in the VM.

For comparison, here are all occurrences of 'iiin' and 'in':
You are not allowed to view links. Register or Login to view.
You are not allowed to view links. Register or Login to view.

Another advantage of introducing additional characters (besides the entropy increase, which may not be substantial), is that the occurrence of 'iii' is more readily explained.

I haven't found more relevant thread to discuss ai-family glyphs, so I'll post it here.
Two ideas concerning ai-glyphs nature:

1) They are vowels and n is an empty symbol. That's pretty obvious. If voynichese is some vowelless language, then we still need vowels sometimes, to specify declension case, etc.

2) (kind of generalization of 1) - they are modifiers for the next glyph. 
Suppose there is default value for each glyph. Take for example r and let the default value be re. Now if we need some similar strings with letter r we define them by attaching some ai-modifer to r: say ar will be aer, air will be ir, aiir - ior.
Another example, say k and the default value is de  Now if we need ad, common suffix prefix in Latin, then define ak to be ad.
Now vowels are produced by taking "empty symbol" (n) and "modifying" it with ai-sequence.
This can explain why ai-sequence can appear in the middle of the word followed by n (then it's a vowel all together) or not followed by n (then it's some modification of the next symbol). Of course there are some a's at the end of the word, but this is a special position where everything changes...

Mary D'Imperio gave this chart at the end of her 1978 work:



I've often wondered if the things that Job mentioned appear in the manuscript: maybe [in] = "um" and [iin] = "unt", so that the result is Latin. But this is only a passing idea - nothing that I have researched

[quote pid='6382' dateline='1474637604']
ThomasCoon:

Mary D'Imperio gave this chart at the end of her 1978 work:



I've often wondered if the things that Job mentioned appear in the manuscript: maybe [in] = "um" and [iin] = "unt", so that the result is Latin. But this is only a passing idea - nothing that I have researched
[/quote]


Thomas, I haven't seen d'Imperio's chart before (I usually avoid reading old theories because I tend to get tunnel-vision when I do) but she's mostly right. Her chart needs some tweaking, as do the letter forms, and it's incomplete, but I can confirm that it's close enough to get the general idea across that many of the VMS glyphs are based on Latin abbreviations.


However, there's no firm evidence yet that the underlying language is Latin. It's possible but... the problem with most of the attempts at Latin decipherment rely on a great deal of subjective juggling of the abbreviations to give them different values to overcome the problem of entropy.

Even in the middle ages, Latin abbreviations had different values in different languages and d'Imperio's chart only describes the Latin extensions (and only a few of them). English, French, German, and Italian scribes applied their own interpretations to the same shapes. Some of them have commonalities with Latin, but some are specific to the language. Also... if the VMS is enciphered, the same shapes may have been applied in a completely different way.

JKP,

I agree on both counts. While trying to save money on paper (which wasn't so cheap in the Middle Ages), scribes needed to save space, which led to a whole complicated system of abbreviations as you mention. And since nobody spoke Latin as a native language (but every scribe spoke it as a second language) the Latin abbreviations could drift into German or Italian or French texts.... thus complicating the situation, if the language does turn out to be a European language