> First of all, Turkish and Mongolian are part of the Altaic
> family too, so we need only leave it at "its appearance in Altaic".
> Second of all, what is the basis for the reconstruction
> of such a sheep word in Altaic? Is the basis a realistic one
> or is it based on a comparison of what are actually late loanwords
> that have entered into the Altaic languages independently? I
> presume the latter. Thoughts?

Glen I have difficulty with the Altaic family, as it seems to be a lot
"loser" in form than do the other big groups (ef, Uralic or
Ural/Yukaghir, Indo-European or Elamo-Dravidian or even Austronesian.)
In this respect it appears in some respect similar to Afro-Asiatic.
It also has similarities to Pama Nyungan in Australia, where
links are seen as long established "trading connections" which
have the effect of making the construction of a single phylogeny
difficult. I know in some respects Mongol is closer to Turkic, and in
others is closer to Manchu. There are also a number of interesting
Manchu-Turkic similarities not shared by Mongol. I wonder whether we
don't have significant substrate sharings under an Nostratic
superstructure. Bomhard devoted a whole section to the problem of
Altaic, and if one includes Korean and Japanese within this family (as
you seem to) this magnifies the problems significantly.

Regarding the presence of a similar word for sheep in these languages,
then I too would suspect that they are travelling loan words.

Regards

John