The Voynich Ninja

This idea has been in my mind for year or so but I am not clever enough to understand nor prove if this could be something that could work.
I have always thought that the glyphs mean more than letters but not gettin nowhere.


Thinking:

Voynich word are about 4 glyphs long, quite short to convey semantic information.
There must be an expansion method, glyphs must pack more semantic punch in them.
Also, in labels the glyphs can stand alone which also could indicate that they are more than bigram-trigram expansions, but whole words.


Idea:

Could Stolfi's core-mantle-crust model, combined with slot system, provide any clue?
Latin is SOV (Subject-Object-Verb), SVO, VSO and poetical mode (OSV,OVS and VOS) structured language, very flexible.
Both Rosam puella amat,Amat puella rosam are acceptable.
Note to reader, I do not know Latin much.

Most deciphrement proposals do not produce grammatically sound language.
If we assume that VMS is based on Latin language, the words and glyph combinations must convey grammatical information, diacritics.

Coming to core-mantle-crust. If core or core+mantle combinations could indicate word and mantle the diacritic how to read the word.
Rosa, rosae, rosam, etc. all from same core+different mantles
Something like Core = semantic content (1-2 glyphs), Crust = semantic modifiers (1-2 glyphs), Mantle = grammatical markers (1-2 glyphs)
Do not take these fixed that core is and mantle is, they can be interchanged. The idea is that we have three components to codify word+grammar, and Stofi's idea has three components.

Then come the slots, if the latin word then depends also on the slot, so even same core+mantle could mean different word in different positions.
There could be then four codebooks:
Book 1: Core meanings for slot positions 1-3
Book 2: Core meanings for slot positions 4-6
Book 3: Core meanings for slot positions 7-9
Book 4: Mantle diacritics and grammar rules

A bit besides the point, but this could possibly give reason for the Currer a and Currer b differences but I haven't given this much of a thought really.
But the idea in my head is that if VMS was made in two different locations, A location and B location, they could have had also two different sets of codebooks

Possible?

Core combinations:
8 basic cores
With crust combinations: ~8 × 10-15 possible crust patterns = 80-120 core+crust stems
Across 8 slot positions: 80-120 × 8 = 640-960 distinct semantic meanings

To write botanical text just from hat:

Plant names: 20-50 words needed
Plant parts: 50-100 (root, leaf, flower, bark, etc.)
Properties: 100-150 (hot, cold, dry, moist, sweet, bitter, etc.)
Actions/effects: 100-150 (heals, causes, prevents, strengthens, etc.)
Medical terms: 100-200
Common words: 200-300 (and, of, with, for, against, etc.)

In my mind this system could plausibly generate grammatical Latin with sufficient vocabulary for botanical description. Enough words could be generated, the grammatical encoding could be enough to handle Latin's complexity.
Yes, there are still things that need explanation, like how we could distinguish words like malum (apple) and malum (evil), how gender would be conveyed, tense (if present at all).

And yes, it would be quite hard system to learn, but just as an idea that I have had in my mind for long and as I do not have the brain capacity to develop this further, so I hope someone here can say the final word for it.

I'm not sure I understand the idea, could you provide a few examples of how this might work? Say, someone wants to encode "Sic transit gloria mundi" using this method, there are some prepared codebooks, etc. What would be the steps?

I am talking only at conceptual level. As I wrote, I do not know how to develop this idea further at the moment.
So I am not saying that this idea is practical, nor saying that this can work. 

It is just a conceptual idea, that maybe combining some known theories or facts of VMS together could produce valid text in some language. Not inventing anything really new, relying on existing research. 

So thequestion that I have been thinking has been, Could Core+mantle+crust with slot theory combined with codebook idea produce grammatically valid language, possibly latin, but at abstract level any language. 

I have no answer, but as we have 8 basic cores, then with crust combinations we could have ~8 × 10-15 possible crust patterns that equals to 80-120 core+crust stems
With 8 slot position we could get 640-960 distinct semantic meanings. 

Not maybe sufficient to write a book but enough to think it as an idea.

I think codebook by itself is more than enough to produce grammatically valid language, either by assigning different codes to different forms, or assigning separate codes to roots and suffixes/prefixes.

As far as I understand, slot grammars (if that's what you are referring to) and core-mantle-crust are two descriptions of the same phenomenon of certain structure in the sequences of characters, so I'm not sure combining them makes any difference. But maybe I'm missing something.

If I understand it correctly, you suggest that words are not coded letter by letter but instead there are some codes for Latin endings.

For example: Gladius - sword

Gladius
Gladii
Gladii
Gladiorum
Gladio
Gladiis
Gladium
Gladios

So you store "Gladiorum" as "Glad" which is core plus "iorum" which is a typical ending and can be saved with some single number.

Well, it is possible. Actually a lot of stuff is possible at the moment.
If you like this idea, follow it and walk that path. Eventually you will learn if you reach dead end or not

Yes, it could be. Actually each 'piece' of a Voynich word (however you define a 'piece') could encode (in however way) some kind of information (including case-endings). The big problem is to find a way to make it work.