To what extent should the translation respect the grammar?

[tavie]

I don't find my question strange, but I will try to rephrase it anyway. If known authors writing in the same languages but in different dialects did not apply the same rules of grammar, we should perhaps be more concerned with the radicals of words, especially those common to both languages, to speed up the overall understanding of the text and leave the refinement of grammar for the next step?

I'm probably not understanding.  I don't think the arguments that some (like me) have made for more grammar in solutions is about refining grammar.  The issue has been that no proposed solution as far as I know has shown any grammar. 

All solutions seem to only provide vocabulary in isolation.  As we've seen, sometimes this can be done to the extent of hundreds of words, in different languages, because it's quite "easy" to read words into Voynichese, particularly if you make a few tweaks.  And because they all do that, none of the proposed solutions manages to distinguish itself from the others.  If it's speed you're after, this approach is counterproductive because in every case so far it has meant someone wasting an awful lot of time and effort on what is a dead end.  

Grammar in the form of coherent sentences is the main indicator that the words found may not be due to coincidence and confirmation bias.  But that doesn't mean we need to assume or expect the grammar to be identical between the scribes.  Instead it makes it more to study them individually.