[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

VMs: Re: RE: VMS evolving table encipherment

Jacques Guy jguy@xxxxxxxxxxxxxxxx wrote
14 December 2003 02:24
Subject: VMs: RE: VMS evolving table encipherment

> 13/12/2003 8:20:45 PM, "John Grove" <John@xxxxxxxxxxxx> wrote:
> >fa   ao   oi   ii   in   ns   sh   ho   ok   ka   ar
> >o C  i O  i U  i L  s Y  h O  o U  k T  a A  i K  r E
> >               n D
> >Doesn't it bother you that 'i' does not exist as a separate entity
> >and is probably just a portion of single character?
> That would not worry me at all. One just has to redefine the
> 'alphabet'.
> Let us have a closer look (this is the first time in all those
> months that we've had a description which would allow me code the
> encryption algorithm).
> When you have generated <ii> you turn to the "ii" substitution
> table to encipher the next plaintext letter. But that next letter
> once enciphered, can only be (from memory) <i>, <r>, <n> and
> <l> (Frogguy: i, 2, v, x -- I'm not fluent in EVA). Doesn't
> look good: far too ambiguous. Solution? Decrete that the
> next Voynich character is a null. So whenever you see <ii>
> you know that the next EVA letter is a null. Fine with me.
> Fine but... it falls down immediately. If this was how
> the VMS was enciphered, you would observe that, for
> every sequence of two characters (except the "null announcers"
> like <ii>) the next one follows the same frequency distribution
> as the letters of the plaintext. And in particular, if the
> plaintext uses a 20-letter alphabet, you will find a similar
> variety in what follows any two-character sequence (except
> of course, the "null announcers").

I agree and this could make thing fall down. If we do assume that iin etc
really single characters it puts a whole new perspective on things. For me
seems to be a way in. It might after all be shorthand as Nick suggests. It
be artificial. For the first time I am able to see some sense in the VMS.

> For instance, following <da> you should see sometimes
> <i>, sometimes <a>, sometimes <o>, sometimes <n> and so
> on, in brief: the whole set of Voynich letters.

This would only occur if the VMS did in fact contain nulls at all.

> Not so at all.  In fact, I cannot think of any two-character
> sequence that occurs followed by even half of the Voynich
> set of letters, let alone the whole of it. And that is
> whether you take <iin> as a single letter, or two, or
> three.

In the early nineties I developed a sort of neural net that stored language
similar to the attempt I made here. I found that patterns did not accumulate
letter, simply because the structure of the language did not allow it. Some
would only
have one letter and some patterns did not appear at all. Surely it depends
upon the
structure of the base language.


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