Have we ruled out simple substitution unwisely?

[oshfdk]

I thought I made it clear that I'm trying to discuss the possibility that transliteration alphabets have potential blind spots in their use.  

Most certainly. I don't think anyone who tried comparing the actual writing with the transliteration would expect the transliteration to capture all the nuances of the script.

However, I think the argument to entropy is not about transliterations, but about solutions. If a person has a solution, the question is how this solution explains the low entropy of the plaintext. If the solver answers that each symbol actually corresponds to 3 different plaintext characters depending on how the strokes overlap (for example), this answers the entropy question immediately. The question about entropy only stays for naive 1-1 mappings of EVA to plaintext, but these never produce any interesting results.

Overall, as far as I can see everything you say is correct, except for the main premise that people are somehow dismissed because of some blind spots in the transliteration. They are dismissed if the solution itself produces a text with unrealistic entropy.