How could we approach the Pharma and Recipes?

[Anton]

However, I am reading regularly that we may not assume that Voynichese words are words as we would think of them. If that is true, vocabulary and pattern comparisons will also fail 

There are some tricky things that can be done beyond simple word frequency analysis. You can look for consistent patterns (think of "so nim").

You can look at positioning. When I analyzed "Voynich stars" several years ago, I undertook what I think was an altogether novel approach: I examined where exactly they do occur in the folio space (of the botanical section), and with this tried to trace the narration structure of the botanical section text. I failed, but with the recent developments it can be explained by that those labels are probably NOT the object names. But the approach is worth of further development.

And so on.