Notes on the Voynich Manuscript - Part 9	[1992 January 8]
----------------------------------------

Curious Parallel Word Lists

Over the past couple of weeks, I've been on a pretty wild goose
chase concerning the vowel system (if any) of the Voynich text,
and another concerning the implication of the ligatured letter
forms.

In this process, a couple of small ideas came up that might be
useful to the group; they concern parallel lists of words that
differ only in their initial letters.  I use the j.guy scheme
of transcription.

The first issue is mentioned by D'Imperio [4.1.1 (13) p 28],
and concerns the mysterious letter "4".  This appears only at
the beginning of words, and only before the letter "o".  What
is more surprising, is that it indeed seems to be a detachable
prefix, since as far as I can tell, every common word in both A
and B sections that begins "4o-" has an equivalent that begins "o-".

From VA:

	4o8aiiv [10]	o8aiiv	[22]
	4olp9	[23]	olp9	[31]
	4olpaiiv[18]	olpaiiv	[25]
	4olpox	[18]	olpox	[24]

One might conjecture that this was a prefix such as "an-" in Greek,
or maybe a contraction of "et-" as in "pepper &salt &garlic", but
if so, why only in front of words beginning with "o-"?

[Note: as a complete scan of the corpus will show, there are in
fact a very, very small number of occurrences of '4' that are not
followed by 'o'.  This warrants further work, I think.]

The second issue is the inflection mark above the ligatured letter
that looks like "cc" with the tops run together.  The two forms are
Currier's "S" and "Z", and Guy's "ct" and "c't".  Again from VA:

	c't9	[51]	ct9	[102]
	c'taiiv	[12]	ctaiiv	[37]
	c'tc9	[32]	ctc9	[54]
	c'tcc9	[18]	ctcc9	[15]
	c'to	[83]	cto	[37]
	c'to89	[12]	cto89	[28]
	c'to8aiiv [10]	cto8aiiv[18]
	c'tox	[92]	ctox	[214]
	c'to2	[65]	cto2	[146]

You know, I find the stability of those relative frequencies astonishing.
Even if the little accent is a modifier of some kind, why should, near
enough, 30% of the words get modified, for every one of 6 words?

In the VB text, we find these letters in other than an initial position:

	oxc'tc89[16]	oxctc89	[30]
	oxc'tc9	[10]	oxctc9	[16]

At present, I have no hypothesis to explain these patterns.  However, one
hypothesis they refute, in my opinion, is that the underlying language is
a synthetic language based on a pure left factoring, or tree structure,
such as the Dewey decimal classification or the Real Character of Wilkins.

In such a scheme, the initial letter is the major category, the next
letter the subcategory, and so on down the tree.  This is likely to produce
"words" that differ in the later letters, but not "words" that differ in
their initial letters.  Thus,

	ctox	Gods, pagan, Egyptian, anthropomorphic
	cto2	Gods, pagan, Egyptian, theriomorphic
but
	c'to2	Plants, European, medicinal, febrifuge

This may be a part of the Voynich "language", but it surely isn't the
largest part.

Finally, if I'm allowed to assume "4thing" as a variant of "thing", then
on f82v (Brumbaugh p104), I claim that of the 14 words that label the
drawings, eight occur also in the text.  (There are actually 15 labels,
but one, "oqpc8ax", occurs twice.)  More food for thought.

Robert