VMs: f78r viewed as columns...

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi everyone,

I've had a look at f78r from the point of view of some kind of four-column 
state machine (an idea various people have been trying out recently): 
here's a link to my RTF...

	http://www.nickpelling.com/voynich/f78r_columns.rtf

Here are my observations...

One possible model for f78r is:-
Column 1:	( )	o	qo	y	ol	or	l
Column 2:	ch	sh	k	t	d	r	s
Column 3:	e	ee	a	o	ckh	cth
Column 4:	dy	ky	y	n	in	iin	iiin	r	ly	m
		(ol	or)

Note that <p> and <f> appear to get added to either Columns 1 or 2, quite 
independently of this model.

One surprise for me was that, while <k> and <t> live in Column 2, <ckh> and 
<cth> seem to live in Column 3.

Also: some same-column pairs appear in either order (ie, <e.ckh> and 
<ckh.e> in Column 3), while others usually appear only in a particular 
order (ie, <t.ch>, <k.ch> in Column 2).

In addition, <s> may have a secondary meaning when used as the first 
(embellished) letter of a line... but that's just my speculation, really.

If you wanted to write a simple state machine to generate (or even hoax) 
f78r-like Voynichese, you might therefore start with a rudimentary 
algorithm something like this (pausing only to season the frequency 
distribution to taste):-

Col 1
Choose from:	( )	qo	y	ol	l	o	or

Col 1	-->	Col 2:-
( )	-->	ch	k	t	sh	d	s
qo/y/ol -->	ch	k	t
l	-->	ch	k
o	-->		k	t
or	-->	ch

Col 2	-->	Col 3:-
k/t	-->	e	ee	a	o
ch/sh	-->	e	ee			ckh	cth
d/s	-->			a	o

Col 3	-->	Col 4:-
e	-->	dy	ky	y	ol	or
ee	-->	dy	ky	y
a/o	-->	n	in	iin	iiin	r	ly	m

Of course, real Voynichese is rather more arbitrary than this simplistic 
model suggests - but many of the differences (<p>s and <f>s aside) may 
simply be down to copying mistakes.

Cheers, .....Nick Pelling.....


______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list