19-02-2023, 12:30 PM
(19-02-2023, 08:04 AM)MarcoP wrote: You are not allowed to view links. Register or Login to view.For instance, dy occurs in ~18% of tokens; there is no language in which "sama" or any other quadgram occurs with such a high frequency.
With all due respect, my proposed scheme is not a way to render information, it is a scheme to render sound. What can you say to repeating 'I' characters with regards to entropy? There are languages like Arabic where vowel sounds do not even appear, so you can only see long consonant chains, and if you transliterated Arabic with only consonants in the target language, you couldn't pronounce anything of it.
And, in case Voynichese is a conlang, which couldn't be dismissed easily, the rules about natural languages wouldn't apply and the composer of the language will have to use a subset of all that's possible in the envisioned language, which would favor some digrams or trigrams or quadgrams way much more than statistically expected.
Just look at a recipe book, and statistically study how many times the word spoon appears. The word spoon would never appear with the same frequency in Wikipedia. The Voynich MS can't be considered for Big Data Analysis in my opinion. Rather, forensics would best suit it.