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

Re: VMs: Number crunching the Fincher window

13/09/2004 10:29:46 PM, Elmar Vogt <elvogt@xxxxxxxxxxx> wrote:

>Here's what I got -- number of different sequences for different sequence 
>Length   VM        German
>4        4389        9435
>5        8773       14949
>6       14087       19623
>7       19432       23443
>8       23934       26264
>9       27263       28237
>10      29527       29609
>11      30954       30543
>12      31783       31187
>13      32249       31651
>14      32491       31964
>15      32612       32190
>16      32674       32346

>This seems to me to point to the hoax theory. If we assume a Fincher window 
>of less than 16 chars width, that would be what we'd expect -- Correlation 
>drops drastically as soon as the window width is reached. I had hoped though 
>for a more pronounced "drop" at the actual window length.

>Or am I completely on the wrong track?

You are comparing with German, which has none of the 
peculiarities of Voynichese (Hawaiian-like low entropy, 
"Chinese-like" word structure etc.). Next, what is the 
German text you used? 

A word list has markedly different repetitive
patterns from an historical account or a piece of
fiction. Part of the VMS is a herbal, another is 
pharmaceutical recipes. Are the properties of such texts
in a known language the same as, say, an historical
narrative in the same language? That would have to
be established first, and in a large number of languages.

The first time I came across a text in Italian (I must
have been eight or nine) I was certain that it could
not be a real language--it had to be gibberish or a hoax.
All the more so that it vaguely looked like French.
Clearly, when running out of imagination, the hoaxer
resorted to French words, disguised by systematically
adding vowels.

Now just think of Chinese. Disreguarding the tones, there
are only 400 possible different syllables, the longest
spelt with five letters (using pinyin, more using
Giles-Wade, more again using the system of the Ecole 
Française d'Extrême-Orient). Then pick a Chinese 
translation of your German text, in each of those
three transliteration systems, with and without tones,
and with a reduced tonal representation. Compare the
statistics. Next, pick another language, and do the
same thing. And another. Try to pick grossly different
languages: German, Chinese, Hawaiian, Basque, Finnish, 
Mingo (an extinct Iroquoian language, very weird), and 
so on. Then and only then will you have the beginning 
of a valid comparison. "The beginning" only. I picked
Mingo because there is an existensive web site about it.
No labials, no "l", no "r", only eight consonants, but
seven vowels, two of them nasals (quoting from memory).
And you can reinterpret /h/ as the aspiration of the
neighbouring consonant, which give you another possible
transliteration, this time with about 15 consonants.
"Well, do it, then" I can hear you telling me. No I won't.
1. It would take time, much time.
2. I don't think I'd ever get it published. Even by
   Scientific American. Rather, especially not by 
   Scientific American, unless I dress it up in
   sufficient bullshit.

Otherwise, you have to elaborate a theory of the properties
of meaningful texts.

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