Alexander:

>IMO, when comparing Uralic & IE versus Uralic & Dravidian we need to >use
>arguments not influenced by our models.

We need models to interpret data. What we do need to do is to develop more
_realistic_ models. The possibility of close relationship between IE and
Uralic is becoming uncontraversial. It is not influenced by "models" but
rather on the grammatical commonalities of the two language groups.
Neglecting models only gives us free license to concoct any sort of
half-assed connections. Models are used in all other theoretical sciences. I
do not see how linguistics is any different.

Alexander:

>I have a great lot of speculations (Like many of us, I believe. Any
>internally incontradictive hypothesis basing on the deficiency in >facts
>available at the moment is a speculation, isn't it? After new >facts arrive
>it becomes either theory or rubbish). But in this case >it was just a
>question, a though aloud.

It's somewhat hard to distinguish "speculation" from "theory" since it's
only a matter of extent and probability. Long-range linguistic comparison is
nothing BUT theory and speculation, so I wouldn't fret too much about it.

At any rate, my personal strategy is Borg-like - continuing approximation. I
continually review my theories looking for errors in logic. These errors are
then corrected to conform to present knowledge. This may cause a cascading
effect, making changes throughout the theory. However, in all, the theory
adapts and evolves, becoming stronger as time goes on. The theory inevitably
becomes more approximate to the truth. Therefore, I am not a slave to an
unflexible model.

- gLeN

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