Notes on the Voynich Manuscript - Part 21	[1995 January 8]
-----------------------------------------

First, a Happy New Year to all members of Team Voynich!  We've been
dormant for quite a while, so how about a resolution: that we'll be
able to wish each other the next HNY in Voynichese?

As a suggestion to restart us: suppose we each write a short note
saying what we're reasonably sure of, based on the story so far.
I'm preparing one on sort of "global" Voynich issues, and as you
might guess my number one is

1. The Voynich Manuscript is not cypher text.

Then we can collate and maybe see where the consensus falls.

Meanwhile:

	Observations on the Orthography of the Voynich Manuscript

These are opinions, not certainties.  But they are, I think,
supportable from the text.

First, the Voynich script was invented.  Its inventor had good
knowledge of contemporary Mediterranean script styles, but made
up most of the Voynich characters afresh.  And they were made up
specifically to allow them easily to be written with a quill pen.

For example, while most modern "joined up letters" cursive scripts
are designed so you lift the pen from the paper as little as possible,
the Voynich script was designed so you lift the pen often.  Look at
the "o" and "a".  That's because you want to make as many downward
strokes as possible, and would prefer two downward half-moons to
one complete circle.  And you also need a lot of places to lift the
pen, because it doesn't hold much ink and needs to be refilled after
every few letters, as you can see from the density in the photocopies.

Secondly, as I think most investigators have discovered, each Voynich
symbol is not a separate letter.  Some letters take two or even three
symbols.  That's one reason our vowel/consonant ratios go wrong, and
also why there seem to be too few consonants.  For instance, if in
English we wrote "m" as "nn", and "w" as "uu", then the actual set
of phonemes would be unchanged, but a statistical analysis of written
text would reveal two fewer consonantal letters, and a biased vowel
to consonant ratio ("mew" would not count as 2C 1V but as 2C 3V).

Here are my specific suspicions.

1.  The symbol "4" stands for 'and'.  But it does not spell the word;
    it is an ideograph with the same function as the "&" of our keyboards.

2.  The letter "o" starts (the most common form of) the definite article.
    But I suspect the full article ends with a consonant that is elided
    in the written text.

3.  Long vowels are represented by reduplication, thus "c c" is long "c".
    This is one reason our vowel counts are off.

4.  Some consonants are made of several symbols, and among them are
    "iv" and "iiv".  This is another reason why both our vowel counts,
    and our lists of consonantal letters, are off.

5.  I believe there is at least one systematically elided consonant
    in the Voynich orthography.  For example, I don't think final "9"
    is just a variant of "a".  I think it stands for "a" plus a final
    consonant.  My current candidates for possible elided consonants
    are L (anywhere), N (anywhere) and S (final).

    As a real example, consider these orthographic changes:

	"sancto" (Latin) => "santo" (Castilian) => "sa~o" (Portuguese)

    Hmm... and wouldn't the spelling rules for Portuguese also work
    pretty well with Catalan?

6.  The letters "cc" (ligatured) and "ct" have some relationship.  I
    once thought they reflected a dialectical difference, like that
    between P- and Q- Celtic, but now I'm not so sure.

And one final point.  If even a little of the above is true, then no way,
no how, could the VMS have been forged by John Dee.

Robert Firth

PS: The gremlins from Qlippoth have been haunting my dreams again.  Their
latest suggestion is that the VMS originated in the city of Srinagar, in
the Vale of Kashmir.  This time, I don't believe them.  But that did
remind me of the Sanskrit orthography, with its systematic elision of
the vowel "a".