Elmar Vogt just posted up some nice statistical analyses of the Voynich Manuscript’s language, looking particularly at the problematic issue of line-related structure.
You see, if Voynichese is no more than a ‘simple language’ (however lost, obscure and/or artificial), there would surely be no obvious reason for words at the beginning or end of any line to show any significant differences from words in the middle of the line. And yet they do: line-initial words are slightly longer (about a character), second words are slightly shorter, while line-terminal words are slightly shorter than the average (though some of Elmar’s graphs get a bit snarled up in noise mapping this last case).
The things I infer from such line-structure observations are
(a) any fundamental asymmetry means that Voynichese can’t be a simple language, because simple languages are uniform & symmetrical
(b) it’s very probably not a complex language either, because no complex language I’ve ever seen has done this kind of thing either
(c) the first “extra” letter on the first word is either a null or performs some kind of additional function (such as a vertical “Neal key”, a notion suggested by Philip Neal many years ago)
(d) the missing letter in the second word is probably removed to balance the extra letter in the first word, i.e. to retain the original text layout, while
(e) the last word has its own statistics completely because words in the plaintext were probably split across line-ends.
In Voynichese, we see the EVA letter combination ‘-am’ predominantly at the right-hand end of lines, which has given rise to the long-standing suspicion that this might encipher a hyphen character, or a rare character (say ‘X’) appropriated to use as a hyphen character. For what conceivable kind of character would have a preference for appearing at the end of a line? In fact, the more you think about this, the stronger the likelihood that this is indeed a hyphen becomes.
But there’s an extraordinary bit of misinformation you have to dodge here: the Wikipedia page on the hyphen asserts (wrongly) that the first noted use of a hyphen in this way was with Johannes Gutenberg in 1455 with his 42-line-per-page Bible. According to this nice post, “Gutenberg’s hyphen was a short, double line, inclined to the right at a sixty degree angle”, like this:-
In fact, Gutenberg was straightforwardly emulating existing scribal practices: according to this lengthy online discussion, the double stroke hyphen was most common in the 15th century, single-stroke hyphens were certainly in use in 13th century French manuscripts (if not earlier), and that both ultimately derive from the maqaf in Hebrew manuscripts that was in use “by the end of the first millennium AD”.
So if you think Voynichese line-terminal ‘-am’ does encipher a hyphen, the original glyph as written was probably a double-stroke hyphen: moreover, I’d predict that Voynich pages containing many ‘-am’s were probably enciphered from pages that had a ruled right-hand line that the plaintext’s scribe kept bumping into! Something to think about!