VMs: Strong / GC / Welnicki / Jeff / etc

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi Matt,

At 12:42 09/12/2003 -0800, Matt Welnicki wrote:
> If I recall correctly, just before leaving the list GC
> commented that he could "read" several pages (34) of
> the VMS and challenged Jeff to a wager that he could do
> so.  While GC never formally detailed his approach in
> one place, a review of posts and LC Strong's work can
> give some very good insight.

BTW, have you reviewed Strong's worksheets? Strong pored over two pages of 
the VMs, looking for a polyalphabetic cipher - and (surprise, surprise) was 
able to find one (but only by interpreting the characters he saw in an 
occasionally hard-to-reproduce way, and the "plaintext" produced in a 
somewhat artificial way too). Does that mean there really was a pattern 
there, or that he projected what he wanted to see onto the VMs?

Yes, I too have tried to reproduce Strong's "double reverse system of 
arithmetic progressions of a multiple alphabet". If the process involves 
anything like what is on his worksheets (say, work-10 through work-13), it 
involved writing down 7 possibilities (each one taken from a different 
alphabet, shifted by Strong's "13579..." pattern) for each letter and 
choosing the English word that arises from them. For example,

        B E U G H
        X U Y F B
        O A C O S
        E X P T F
        S G L X -
        R C M W L
        H M S P -

...read "EXCEL" (9 9 4 5 11) to Strong (even though "E" fails to appear in 
his fourth column).  However, "HUMPS", "HASPS", "RUMPS", "SUMPS", "RAMPS", 
"HEMPS", etc would seem to be equally (if not more so!) valid 
possibilities. It should be clear that (in this case) he had 7*7*7*7*7 (ie, 
16807) permutations to choose from there. Even so, Strong also didn't feel 
obliged to impose a restriction of consistent spelling on the plaintext 
(COITUS and COTUS, CIFER and CIPHER all appear in his notes). Many words 
seem rather impressionistic in their interpretation of English. Syntax was 
also an optional extra, rarely making its presence known in the plaintext. 
Finally, some claimed words (like "OKRA", IIRC?) were anachronistic (and 
hence almost certainly wrong).

IMO, claimed decrypts of the VMs tend to have three components:
(1) interpretative ambiguity in the process between Voynichese and 
plaintexts (plural)
(2) interpretative choice (between possible plaintexts) in the eyes of the 
decryptor
(3) interpretative choice of meaning (from adjacent words) in the mind of 
the decryptor

The final "meaning" then arises from the combination of all three parts: it 
might be an interesting exercise for someone to assess the proportions of 
these three for all previously claimed decrypts. :-/

In Strong's case, there was considerable processual ambiguity, in that his 
process could generate many possible plaintexts (16807 of them for a 
5-letter word). Then he had considerable scope for choosing English-like 
(misspelt or distorted) words from those. Finally, when all the chosen 
word-like entities get placed together in a sequence you would be 
hard-pressed to recognise it as "English" of any variety, were it not that 
Strong was telling you that's what it was. Yet these are his claimed decrypts.

What Strong did is broadly analogous to what Jeff is doing: Strong focused 
solely on the statistics of the transcription without taking on board what 
the rest of the text was saying... ie, that there is considerable internal 
structure to Voynichese. Really, I'd say that any decrypt that doesn't 
explain why "qo" and "dy" appear together, or why 
pages/paragraphs/lines/words have the structure that they do (to at least 
some degree) is highly likely to be wrong. The statistical signatures of 
the internal structures remains fairly consistent throughout - why should 
that be? Is that purely chance?

GC has often expressed the opinion that these are carefully hand-crafted 
outward signs, grafted on (through careful choice of spelling etc) to a 
Strong-like system - ie, that the system is essentially polyalphabetic (in 
the way Strong described) but that it was made to look as though it had 
deep structure. Well... that's his opinion, not mine.

I think it's a misleading 15th Century cipher system (and that the 
statistical signatures we see are expressions of the set of multiple 
steganographic mechanisms it's using to hide its content) - GC thinks 
(following Strong) it's a 16th polyalphabetic cipher, made to look like a 
15th Century monoalphabetic cipher.

As the man says, "yer pays yer money, yer takes yer choice". :-)

> He [GC] did also mention pizza alot too.   Mmmmm...pizza.

It's not just GC, we're *all* hungry for pizza here... :-)

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


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