Re: VMs: The VMS "alphabet" and some other issues...

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi everyone,

At 13:26 02/02/03 -0500, Larry Roux wrote:
> I also believe that many Voynich characters are actually multiplets. (ie
> t may be "ir", "ar", "tt" or "ch" etc).

I think you're basically getting on the right track here - note that you 
can quickly get a feel for how a page of VMS text could "pairify" into 
appropriate pairs simply by "trying it out".

As an example, I've thrown together a quick demo based how I think this 
might work for the (quite clear, quite small) VMS text on page f38r, in 
this (145K) GIF file:-

        http://www.nickpelling.com/voynich/pairs.gif

The things I generally infer from such "pairification" exercises are:-
(1) "<gallows>", "o <gallows>", and "y <gallows>" code for different tokens
(2) "qo" "ol", "or", "cho", "ar", "on", "an", "os", "as" look like 
different tokens
(3) "y" usually appears on its own (though "dy" might well be a token)
(4) "od" also looks like a token (but the two letters can get split apart)

Observation #1: if "od" is a token, one possible reason for splitting it up 
might be to hide the start of the next word. In fact, I've often noticed 
"obvious" pairs getting split up in other pages - this might be a common 
mechanism. Just a thought. :-)

Observation #2: on the bottom line, the second "o" appears to be dotted... 
and it looks to me as though this could be because, within the coding 
system, there are two possible ways it could make sense pairified (ie, 
either pair it with the preceding "Sh" glyph or with the following gallows 
glyph). The dot might then be an indication as to which way the pairing 
should go...

( ...or it could just be an inkstain. You choose! :-o )

Observation #3: once you've pairified a page in this kind of way, there 
don't seem to be *that* much remaining - ie, it seems to fit the "internal 
structure" quite well (though rather differently from "core-mantle-crust" etc).

However, one problem is that you quickly start noticing non-rule-like 
clusters, which tend to be based around EVA "c" characters. These can be 
difficult to group satisfactorily into a pair structure - looking again at 
the pair of "y"s one above the other near the start of lines 4 and 5, I 
think I've pairified that bit of line 4 wrongly... easy to do. Ah well. :-/

It may be this (weakly) indicates that "c" characters correspond to some 
tricksy method of numbering - but it's hard to be sure. :-(

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

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