As far as I'm aware, there really aren't any firm theories, it's just a completely baffling artifact. Based on dating, it's (probably) not a forgery. It's an incredible amount of effort for the time to be just some person's fancy. There just doesn't seem to be even a theory that fits all the evidence. It's truly bizarre. I love it.
It's a lot of work, but not particularly more work than a scribe would usually do.
People have always been people, we've always been creative and intelligent, and I really think the best explanation is that this is just an odd creative work. We'll never know the exact details.
The leading theory seems to be that it's meaningless gibberish, in part because asserting that it isn't requires some degree of specificity as to what it could be, and there's little agreement on that aspect. Which is to say, it's the plurality theory but far from the majority theory.
It's also known that the text makes no sense as an encipherment known in the 17th century of any known natural language. Statistical analysis of features suggests it's not random gibberish... but also distinguish it from natural language. It's also prone to a high degree of repetition or near-repetition (think phrases like "burgle bugle bugie", with one-letter changes between successive words).
Personally, I think it's gibberish that was constructed to have some degree of plausibility for a would-be cryptanalyst. Or maybe even something as prosaically simple as calligraphy practice, given the unnaturally repetitive nature of a lot of the text.
> some degree of plausibility for a would-be cryptanalyst
But that's a very ahistorical theory, isn't it? The manuscript was written hundreds of years before cryptanalysts were invented. For its author to anticipate that would have required an impossible degree of foresight. Rather, whatever features give it cryptographically interesting properties must have been motivated by something other than a desire to confuse/intrigue cryptanalysts, with that just being a coincidental side-effect.
Morse, ASCII and military phonetic transcribed literally (just to cite a non-machine one) all look like repetitive gibberish if you don't know what to look for.
Morse, ASCII, and military phonetic are all pretty recognizable as single-unit substitution ciphers. We've pretty much ruled out any possibility of it being a single-unit substitution cipher of any known language.
One theory I never seen considered is that the text means nothing, it's a distraction or later addition unrelated or loosely related to the main content, which is all on the pictures.
In other words, the pictures are the original thing and main content. I would guess they're not about bothanical stuff, but instead use names of bothanicals and other alchemical ideas from the era it was written.
My guess is based on the idea that plant names and alchemical ideas last a long ass time. They're perfect for encoding something meant to not be changed by common text-altering techniques (adding punctuation or accents to non-puncuated or non-accented languages, deriving letter-casing from non-cased letters and so on).
The last papers I read on the subject clearly stated that it appears to be a real language given the word and character distributions but it can only be a tonal language, like Sino-Tibetan or Meso-American.
Which in my opinion makes the old Marco Polo delegate theories the most probable, but all through the lenses of European culture.
A very plausible argument I read a while back was: it is a manual of exercise and herbology and related things. The symbols in the book aren't in any language. They are like rough music notation; XXX, YY, XX etc. They are custom for the client of whoever wrote it, meaning "stretch, stretch..., bend, bend..." or something similar.
The theory isn't popular because it doesn't reveal anything but once you look at the book it seems likely. The notably thing is the symbols don't have the variation of a language - they are less complex and more locally repetitive.
