Unusual Words

[Mark Knowles]

If text can be split into meaningless and meaningful then the best that one can do is to start to separate these two types of text as well as one can even if one isn't able to perfectly separate them at the start. The problem one has is that the meaningful text could use a very simple substitution cipher or be very complicated. The simpler the encoding of the meaningful text the easier it would be to decipher it and therefore prove its existence once some of it has been isolated.

[Mark Knowles]

In fact it seems to me that the overwhelming weight of Voynich research has been looking at the patterns and structures of those words that I see as fillers not the unusual words that I am most interested in which I think are often discounted as though they were just background noise.

How can we prove which scenario is closest to the truth?

I think to me what seems to be the best way to do this is to identify and isolate Voynichese text that stands out as unusual when compared with most other Voynichese text. Then the larger the collection of the most unusual text that I can find the better. My inclination is then to study that text in isolation looking a glyph frequencies and other simple measurements with the hope that the underlying encoding is not too complicated.

I have discussed the potential use of string functions to more precisely separate the two types of text. Which function is best is not completely clear to me yet.

[Mark Knowles]

When looking for the non-filler text then unusual atypical glyph sequences are certainly worth looking out for.

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

[Mark Knowles]

In Torsten's work he identities word clusters; I tend to associate those clusters with filler text. So I suppose I am most interested in unclustered words/text.

It is by better recognising and spotting the patterns in the filler text that one can better spot text that doesn't conform to those patterns and so is not filler.

I suppose that I need to better familiarise myself with the patterns in Voynichese, so as to better collect examples of text that does not follow those patterns and so is not filler.

[farmerjohn]

Words with 3 gallows:

Code:

http://daiin.net/browser/search.html?mode=web&pattern=.*([tkpf].*){3}

(Copy-paste the URL, the forum does not allow a link for some reason.)

The link contains square brackets that conflict with BB tags.
I have updated the search page, so the browser address bar now is synchronised with the search string. Thus you can just copy the address and insert it in You are not allowed to view links. Register or Login to view. (somehow the copying correctly manages all the special symbols).

[Ruby Novacna]

Mark, if you can separate the words into two sets, that's a good thing. 
However, I have the impression that your list contains words from two languages, A and B. 
Do you believe that the difference between the two is not important?

[Mark Knowles]

Mark, if you can separate the words into two sets, that's a good thing. 
However, I have the impression that your list contains words from two languages, A and B. 
Do you believe that the difference between the two is not important?

I don't know. However I suspect that the filler text may vary as suggest i.e. filler A and filler B. I would imagine/hope the non-filler text to be a constant language.

[Mark Knowles]

chekolkociph

From the Rosettes folio looks unusual.

[Mark Knowles]

Given concerns about determining spaces and word divisions I wonder if I should have called this thread "Unusual Text Strings". However I should emphasise that by "unusual" I do not mean marginalia and other writing by later individuals on the manuscript; rather I mean writing by the original scribes with atypical glyph sequences.

Of course defining what one means by "unusual" is difficult. I suppose here is where the subject of string comparison functions comes in again->

You are not allowed to view links. Register or Login to view.

[Juan_Sali]

What do you think of the labels? 
Do they have meningless text? 
In some cases they are quite short,  with meaningless text they will be even shorter. If codified, will they be long enough to have a meaning?