Re: VMs: Voynich and bee-dance

[Jacques Guy <jguy@xxxxxxxxxxxxxxxx>]

27/11/2004 11:58:00 PM, Nick Pelling <nickpelling@xxxxxxxxxxxxxxx> wrote:


>AIUI, Dr Paijmans' paper asks whether we can transcribe/notate bee dances 
>so as to look for Zipf-Law-like behaviour, & hence to see whether it has 
>language-like behaviour.

Which is perfectly frivolous since he writes:


"Later other researchers such as Wentian Li [10] 
have proven that Zipf?s law also holds for less 
interesting phenomena, such as randomly generated 
sequences of characters."

So he knows and acknowledges that a Zipf 
distribution has no predictive value. Now
what is the point of his paper then? C'est
du bouche-trou, rien d'autre.




He also writes:

"The consensus of all researchers is that 
the emergence of a Zipf relation between phenomena
in a language-like construct does not prove that 
the item under consideration is a language, but 
that a real langua[ge] almost certainly displays 
Zipf behaviour."

I take exception to that. Show me a study of
sign length and sign frequency in a Sign Language.
(I suspect that Sign Languages do not follow
Zipf's law). At any rate, some of the examples
in that paper are clearly false, such as, for
instance, the populations of cities. Try that in
Germany! Zipf's law holds... for cherry-picked data.

>In fact, the key issue skirted by Paijmans' paper is how one should best 
>transcribe bee choreography given that we don't actually understand how its 
>mechanisms works - trying to recast the problem in terms of "information 
>content" (in much the same way that Gordon Rugg does) is actually quite 
>unhelpful.

It's yet another one of those papers without 
any merit. The field of artificial intelligence 
is full of this sort of bum fodder.



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