Re: [tied] Re: 'Dog' revisited

01-12-03 17:35, ehlsmith wrote:

> Then the argument is even weaker- instead of comparing the number of
> supposed cognates with a small number of language families then it
> has to be compared with over 5000 languages. This just makes it more
> likely the "cognates" are pure chance. And what criteria do they use
> to determine a semantic match? If, as I suggested above, they include
> names for particular breeds then their examples are plucked from an
> even bigger pool

And what if they even include an occasional wolf or hyaena? (which of
course means that the actual word for 'dog' in the language in question
must be something different and the "comparatist" is clutching at straws).

> [Heck, if they even allow for the accural of an
> occasional prefix over the millenia, then they can include coCKER
> spaniel in their "cognates"]

Perhaps an expressive reduplication (*ku-kur-), and in any case bearing
an amazing similarity to Indic (substratal) kurkura- '(one of the words
for a) dog'. :-)

Piotr