No text, but a visual code

[oshfdk]

Unattached endings certainly exist everywhere, including after interruptions, but they do not appear to have been split (we can't be sure, but there is no good evidence) when the re-assembled vord does not exist or the two parts exist elsewhere as two separate vords.

Let me try to rephrase to check that I understand your argument. It went something like this:

oshfdk> <...> if this herbal does split the words using the stem, there is no reason to believe this was impossible in the Voynich manuscript

nablator> There is. <...> Unattached endings certainly exist everywhere, including after interruptions, but they do not appear to have been split (we can't be sure, but there is no good evidence) when the re-assembled vord does not exist or the two parts exist elsewhere as two separate vords.


So, as I understand it, you are saying there is reason to believe it was impossible for Voynichese to split the words using the stem, because it never splits words in a way that produces pieces not seen on their own (say, splitting a||iin or e8||y). And also if you try to reassemble these words, they either don't exist in the MS or exist as separate words. Is this correct?

Naturally, when people split words for one reason or the other (say, hyphenation), they tend to split at certain points, like syllable boundaries. For example, if you want to split "existing", you can do it as "exis-ting" or "exist-ing", it's highly unlikely you will do "existi-ng". Or you can split "carrot" as "car-rot", and less likely as "carr-ot" or "carro-t", and as you can see the default split produces two parts that work as quite common words on their own. It's natural property of most languages that words are built with a limited number of syllables that are both the best boundary for splitting the word and are separate words on their own. 

Also, it's possible that the symbols on either side of the split are some kinds of special marks or have some special meaning, or they are a special (hyphen-like) form of other symbols. Three EVA 's' preceding the split on 26r could each function as some kind of interruption mark. So the reassembled words do not exist, because the form (or presence) of these symbols located next to the split point may change.

Overall I see you argument as more of "what is more likely" than "what is impossible", but maybe I missed something.