Re: [tied] Re: About methodology

From: Glen Gordon
Message: 3471
Date: 2000-08-29

Glen (me) wrote:
> >My understanding is that comparative linguistics is purely
> >theoretical and therfore _inexact_.

Nemo:
>If mathematics in your opinion is "tangible" then I'm not
>surprised that comparative linguistics is "purely
>theoretical". But why "purely theoretical" sciences must
>be necessarily "_inexact_"?

If something is only theoretical it naturally implies that the actual answer
can only be guessed at. Even if that guess is 99% accurate, let's say, it is
still somewhat inexact. There will never be physical proof of prehistoric
languages (unless writing goes back a million years??). When time machines
are built, comparative linguistics will finally become a "tangible" science
that can be physically proved but until then, we have to keep guessing.

> >Due to this, it is quite impossible to measure CompLx
> >with _exact_ tools such as statistics, probability, etc since
> >they are contrary to the inexact nature of CompLx. As well,
> >comparative linguistics involves almost random human and
> >social behaviour in the end.
>
>Statistics and probability are quite good at measuring random
>phenomena.

But not in CompLx because it's very easy to mis-measure valid but obscure
sound correspondances as "random" when they are quite regular. You can't
accurately label human or social behaviour as "random" or "non-random" and
sound changes are mostly that - human behaviour.

> >So, if you get a chance Nemo, I want you to translate human
> >and social behaviour to me into an exact mathematical model.
>
>Indeed, so far there are only _inexact_ mathematical models
>in social sciences. But sometimes it's better than nothing.

This is my point, but Piotr disagrees.

> >When you've done that, we'll get to work on finally measuring
> >comparative linguistics and after that, we'll start devising ways
> >to measure other things like "love" and exactly how many angels
> >there are on the head of a pin. :)
>
>Mathematical models are neutral. There is no reason a priori why
>"love" (or rather some of its aspects) couldn't be
>modelled mathematically. It's only a question whether any model
>of love would be correct, complete, isomorphic etc. and, above all,
>useful. Most probably not, but you won't know for sure if you
>don't try. At least it could be a mental exercise, like those
>angels on the head of a pin in scholastic times.

I make it a point not to waste my time trying to solve logically intractable
solutions. :)

- gLeN

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

Share information about yourself, create your own public profile at
http://profiles.msn.com