RE: VMs: Entropy, was WAR against EVA

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi Larry,

At 07:57 04/03/03 -0500, Larry Roux wrote:
> There are Voynich glyphs that never appear next to each other.  The 
> question is: is it due to the encryption scheme?  Or is it due to the 
> author running an "e" into "i" to make "ch"?

This is only one of the many features which differentiate Voynichese 
(however you stroke it) from other languages. Even if you think you have an 
answer that explains one factor, it is barely the smallest of stepping 
stone towards an integrated/holistic answer which addresses them all.

> In either case, I would think that if this was a hoax than you would find 
> a much more random occurrence of characters and would find ql more 
> often.  Obviously the author had a strict set of rules which were followed.

This is, in essence, the core of the argument against more complex 
enciphering or encoding schemes (such as polyalphabetic substitution) which 
would - in almost all circumstances - tend to "flatten the stats". That is, 
the various distribution curves for the result of a "straight" polyalpha 
encoding wouldn't be so enormously "peaky" as the ones we see - all the 
individual symbols would share a similar probability of occurring.

Code-makers and code-breakers pre-1550 understood (some intuitively, some 
[like Cicco Simonetta] explicitly) that it was frequency distribution that 
usually gives the game away. Polyalpha (in normal use) flattens that 
distribution, which removes many of the obvious ways in.

The main reason that polyalpha didn't take over from "simple" ciphers (many 
of which were not actually that simple) seems to have been due to a general 
emotional attachment to the whole culture that had sprung up around 
ciphering 1350-1500 - polyalpha was better, but people liked the old way 
better.

To claim - as GC does, following Leonell Strong's alleged decryption in the 
1940s - that the VMS' code is a tricky polyalpha, perhaps with the keys 
hand-picked for every page, would almost certainly require the later 
dating, as claimed (say, the 1560s in England).

However, when I look at the 9-rosette page, I see Milan circa 1460: when I 
look at the cipherbet, I see resonant echoes of several tricky Northern 
Italian ciphers circa 1440-1460: the majolica designs appear geometric 
(which points to NE Italy pre-1480), etc: and so I draw the general date 
and place conclusions you'd expect.

Polyalpha would always tend to destroy low-level structure (while 
introducing its own subtle cycle-length level of structure), and it would 
require an extraordinary act of will not only to retain structure with a 
polyalpha, but also to get it to the level where we can observe 
word-centric paradigms (such as core-mantle-crust) or adjacency-centric 
rules being consistently followed.

Strong believed this to be the case: but it simply doesn't gel with my 
reading of the art historical evidence, while also failing to fit my 
understanding of basic statistics.

If the dating is pre-1480, then I believe the code would need to have been 
a composite of several simpler codes/ciphers well-known by that time - 
ciphers, codes, Roman numerals, shorthand, steganography, etc. However, the 
structure within which these separate mechanisms are combined may be the 
root cause of many of the "strict set of rules" you mention.

One of my long-held suspicions is that the VMS may simply be an obfuscated 
shorthand, or (more precisely) an obfuscated tachygraphic ("fast writing") 
system. The basic shorthand system (as evidenced by its single-stroke 
alphabet formation) its alphabet is based on appears to have been designed 
for writing on wax tablets.

But given a small (but fast) alphabet without vast numbers of Tironian 
notae - specialist symbols for frequently occurring words, like <8> 
"cum-"/"con-" , or <9> "-us" - to help, how could a tachygrapher keep up 
with the flow of speech?

The only recorded small-alphabet system that precedes Dr Timothie Bright's 
"characterie" of the 1560s is from a Mr Ratcliffe of Bristol (who was using 
it in the early 1500s), which simply involved dropping vowels and other 
unnecessary parts of speech, but used normal English letters for the rest. 
However, that's exactly as far back as the history of shorthand goes - the 
rest has yet to be found out.

It's a perfectly reasonable hypothesis to propose that - especially given 
its alphabet structure - the VMS is based on a wax-tablet shorthand system 
parallel to Ratcliffe's, constructed in the period between the death of 
notae and the birth of characterie.

I think that a system built around a structured (though obfuscated), 
largely consonant-based, shorthand-driven text would have many of the 
features we see in the VMS. Though YMMV! :-)

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

PS: if someone were constructing a universal language circa 1500, why on 
earth would they construct it to look just like a simple cipher (without 
actually being one)? :-9

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