Usefulness of categorizing words based on occurances in different parts of the MS?

[joben]

I was wondering if this method is feasable:

The MS contains 6 different "parts", Herbal, Astronomical, Biological, Cosmological, Pharmaceutical and the Recipe part.

...

I know these are the historical designations, but I don't like them. If something is bad, I think it's a good idea to change it.

We don't know if the plants are all herbs (some might be trees). I prefer to call them the Big Plants and Small Plants sections (which I suppose is also a bit ambiguous since it doesn't refer to the size of the plant, it refers to the size of the plant drawing, but it's better than assuming they are herbs).

We don't know if there is a Pharmaceutical or Recipe section. The section at the end might be 1) proverbs or 2) good and bad days or 3) historically important events or 4) something else... I prefer to call the two sections of unillustrated text the Dense Text pages and the Starred Text pages, but there might be better designations.

The biological section might be biological or it might be mythological but I can't think of a better name than biological, but perhaps someone else can.


The name "cosmological section" is not as objectional since medieval cosmology was a broad grab-bag of beliefs about the universe (and there are a lot of stars and suns in the images), so I have fewer objections to this name than to some of the others.


Mainly I dislike herbal, pharmaceutical, and recipes. Even if these names turn out to be correct, it's really not good to call them that in advance. If the dense text consists of good and bad days, historical events, a genealogy, a textual calendar, or a list of stones and their properties... something along those lines, then they really don't fit the definition of recipes.

This is a good point and I assume that if I dig deeper into this MS I will start noticing things like this. 

I always assumed the part names was an oversimplication for an easier overview of the MS. 

I think that storing every word in a database together with the number of occurences on each page is a good start. Then if we made it simple for anyone to create their own part names and connect each page to the part they feel it belongs to, then we would get a statistical overview that doesn't use the historical "made up" part names. If this hasn't been done already, I think it would be a simple task for a programmer.