From: Arnaud Fournet
Message: 61647
Date: 2008-11-15
----- Original Message -----
From: "etherman23" <etherman23@...>
To: <cybalist@yahoogroups.com>
>> How reliable are reconstructions of non-IE families?
>> I realize there are 3 or so reconstructions of Afro-Asiatic, all by
> respected scholars. But how do they compare with IE reconstruction?
>> Also Uralic et al. Do we really have anything truly well elaborated
> to compare with yet?
>>
>
> Uralic is very well elaborated. There are open issues, of course, but
> that's no different from PIE or any other language family.
=======
As i have pointed at before,
http://tech.groups.yahoo.com/group/cybalist/message/55999
Proto-Uralic reconstruction is bad.
There are no consistent phonetic correspondences
and most "reconstructoids" are more rough approximations of lexical data
than real reconstructions the way PIE is reconstructed, with one to one
matches.
You have two kinds of reconstructions :
1. the old approximative one you can find in Redei's UEW, (just bad)
2. the new one by Sammallahti and other Finnish people, (much worse)
There are a lot of oddities in those reconstructions,
For example, labial stops like -b and -p never appear in Cv_C position,
Etc.
Plenty of words are not even listed,
Loanwords with Indic phonetics are not even distinguished from those with
Iranic phonetics...
In a way, old works from Cöllinder are better than more recent works !!
At least, Cöllinder looked at the data and was not dogmatic !!
Arnaud