What are labels?

[Mark Knowles]

Namely that the Voynich contains "filler" text which conforms to a certain structure.

Unfortunately, I'm not closely following all discussions in the Forum these days, so I'm not acquainted with this theory of yours. However, it would be strange to have "null" labels, would not it? Consider the "Voynich stars" (f68r1 and f68r2). How many of those would be null labels, according to your theory? f68r2 has some stars unlabeled, instead of null-labeled. Why? You can't but notice that this immediately shifts away from the "best odds" approach. Actually, the null-label theory in itself is not "best-odds", I guess  

I think the Voynich stars is an excellent example. Some stars have no labels, therefore it is not necessary for every star to have a meaningful label. The reason for having meaningless labels is to confuse the person trying to break the cipher, so in ciphers the reason to have "null" symbols/characters is to confuse the person trying to break the cipher as to what are "real" symbols/characters and what not. So "meaningless" star labels would be there to confuse the person trying to break the cipher as to what are the "real" labels. Counting and distinguishing "null" words from "real" words becomes hard and if I could perfectly partition the set at this time that would be great. However, generally, I view rare and distinctly spelled words as being much more likely to be "real" words than very common words with spellings very similar to other very common words. Note the repeated labels tend to be spelled very similarly to each other, whereas the "unique" labels tend to be spelled much more differently from one another; this should be statistically verifiable, which is on my list to do.

In some ways I agree that it shifts away from the simple best odds approach as we can't find any contemporary cipher which contains "null" words. However it is the scenario that seems most likely or the best odds in this context to me, personally.