turkicresearch.com

[-JKP-]

...
So are we expecting a constructed language to be the answer? Or meaningless text? I believe it is unfair to say that the solution should explain the frequency numbers. All languages have a number of linguistic phenomena underlying it. For example as someone pointed out, in English we have multi to multi correspondence for letter/letter combos to sound. But if someone proposes such a solution, we ignore it saying it has many degrees of freedom. I dont think English would ever be reconstructed by people with this mindset in the future if it were to be a lost language at the time.

I didn't say any solution should explain the disparity in letter frequency. There are ways the text might be manipulated to come out like Voynich text. Sadly, the people who offer substitution solutions do not explain this discrepancy in their methods, they don't even acknowledge it. I daresay many of them haven't even noticed it because they cherry-pick the words that seem to work and ignore all the rest (in other words, they ignore about 85% of the manuscript).


What I was pointing out, since this thread is specifically about a substitution-style solution on turkicresearch.com is that anyone proposing a simple substitution cipher MUST explain the positional and frequency anomalies that differentiate their substitution solution from natural language. There is no natural language that has the positional characteristics of Voynich text. Even syllabic Asian languages do not function this way.

That's why all the people trying to do it this way (at least the ones I've seen so far) stall after picking out about 5% of the words that seem to work with their system. The next step, for those who keep going, is to start proposing that it is polyglot, or that each VMS glyph can represent multiple letters. Then they can maybe get up to 10% or 15% "hits" using their cherry-picking method. Applying their method to the rest of the text results in gibberish. In other words, they are SELECTING what works and ignoring what doesn't (which is easy to do when there are 200 pages of text) while simultaneously ignoring questions about the lack of grammar.