Re: [cybalist] Re: Glottochronology.

From: Dennis Poulter
Message: 2214
Date: 2000-04-27

Thanks Andrew, for your posting.
I wasn't seriously suggesting that language can be modelled using Chaos
Theory, even though the characteristics of language change seem to have much
more in common with systems that Chaos Theory claims to describe, rather
than mechanistic models derived from the natural sciences.
As for statistical analyses, I don't think we have got a large enough
perspective yet to iron out the irregularities. After all, there are only
some 200 generations between PIE and modern English, French, German,
Russian, Farsi, Bengali etc.
Added to this, languages do not only diverge. They also converge. There are
also "areal features", developments which do not seem to respect
conventional language groupings.
I think that, while on the surface mathematical analyses appear to promise
much in explanatory power, this is illusory, and underneath they are
seriously flawed and misleading. Furthermore, I don't think that
mathematical modelling will ever be viable for language - people and logic
do not mix.
Obviously, there are many serious scholars (who receive serious funding) who
would not agree with me. Maybe I'm just an old Luddite, but I don't like the
idea of reducing language to numbers, or translating machines or talking
computers for that matter. Language to me is such a uniquely human
accomplishment, and I love it for its beautiful illogicality. I can only
echo the cry of the Prisoner in the old sixties TV series : "I am a man, not
a number".

Cheers
Dennis