The incompatibility of Voynichese with natural human language

[Alin_J]

Thank you, Jonas!
As a percentage, I consider the number of perfectly reduplicating consecutive word couples vs the total number of consecutive word couples (N.words-1).
According to the numbers you provided, we have 64/40.000 i.e. 0.16%, about one fifth of the VMS average. But, as pointed out by your expected/actual argument, it is clear that these occurrences are very significant.

It seems to me that the majority of the cases occur in the position you described: around line breaks (a line ending with word W, followed by a line starting with word W). This is something I have never seen: definitely interesting.
Possibly even more interesting, the text also features several instances of quasi reduplication. E.g.:
tuuriasi, tuumiasi.
mustien mutien
selvältä selältä
My first impression is that, also in this case, we are far below VMS rates, still it is very exciting to see a text featuring both systematic reduplication and quasi-reduplication!
Thank you again for mentioning Kalevala!

Hey, I thought I should also mention now that I happened to just have analysed it, that the Japanese language sample (N-JPN) in You are not allowed to view links. Register or Login to view. contains 454 perfectly reduplicating words among the first 40710 tokens. According to how you measure, this is about 1%, isn't it?

Because the type-token ratio is so low for this text, this is about the same as you would find if the tokens were shuffled randomly, so according to my way of counting it is not as significant as Kalevala in this regard. But it is according to how you count. There are still no higher number of repetitions than pairs though.