To what extent should the translation respect the grammar?

[Koen G]

There are different definitions of grammar at play here. One is spelling: this was variable in medieval texts. Two is "making mistakes", as is expected when the author does not master Latin. Three is "Voynich translator word salad propelled by Google Translate".

I might add to tavie's excellent reply that I remain deeply convinced of the following: if someone finds and publishes the solution to the VM, we will recognise it. We will know that it has been solved. No full translation will be required, just a methodology that addresses the issues, and a small demonstration. People will recognise it. At Yale, here on the forum, elsewhere, it will be proclaimed that a solution has been found. The language or cultural background doesn't matter, be it Turkic or Semitic or Latin or Greek or German or 19th century English. What matters is that the solution recognises and addresses the issues that make simple substitution impossible. If such a solution is demonstrated to work, I can guarantee that it will be recognised. But so far I have only seen isolated words being selected, or word salads, or creative writing exercises. No solutions.