Re: VMs: Arguments against a code book?

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi Elmar,

At 12:34 12/07/2004 +0200, Elmar Vogt wrogt:
> Lately the list discussed the idea of the VM being a code a bit more 
> vehemently
> again. (Ie, every symbol group in the VM might represents a "token" which 
> would
> be looked up into dictionary wordlist.)
>
> Before actually tackling the question how such a code could be broken, I
> pondered:  Most of the theories about the encoding algorithms of the VM fall
> flat quite quickly, because they don't match with the properties of the code
> text. Are there any similar observations which rule out a code book from 
> the start?

This theory's many variants tend to see each word as some kind of 
obfuscated number grouping, whether Roman numbers (as Tiltman discussed), 
or something more exotic - for example, I once suggested that the gallows 
characters might be loosely related to David A. King's "Cipher of the 
monks" (a Cistercian numbering scheme based on lines sprouting off from a 
vertical line): and also once suggested that they might be a kind of 
single-stroke shorthand for X/XX/XXX/XL, based on the number of times the 
single stroke crosses over itself.

If you then extend this to see all Voynichese words as somehow forming 
indices into a code-book, you've got yourself a nice hypothesis to test (as 
Marianna described recently)... except that it's really not clear how it 
could work. Yet again, the devil's in the details.

The problem is that we can't obviously see how to generate Voynichese words 
- and hence we can't really see how they could be a numbering scheme. 
Models (like crust-mantle-core, and the many variations on that general 
theme we've all proposed from time to time) capture ~some~ of what we see - 
but an awful lot remains. Could that huge residue simply be nulls? It seems 
unlikely to me (though YMMV).

Anyway (back to your question), if Voynichese were a simple-minded 
code-book (but perhaps with a fiendish number encoding schema), I think we 
might expect to see (1) rather more obvious matches between labels and text 
(etc) than we do, (2) a language-like distribution of word frequency, and 
(3) uniform text structure across the text (unlike the differences between 
Currier A & B we do see). Something to think about, anyway. :-/

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


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