Glen

> Whether we consider Tungus to be related to Japanese, or Korean
> related to Mongolian, is one thing - a matter of individual
> disagreements on what constitutes a genetic relationship. However,
> there is surely a common core of words within Turkic, Mongolian,
> ManchuTungus, Japanese and Korean that can be called "Proto-Altaic"
> and this language most certainly was spoken somewhere in Central
> Asia in some form or another.

Agreed, but in the absence of a soundly based phylogeny, we will be
tempted into errors in our reconstruction of Altaic (a little like
linguists of the 19th century made errors by accepting Sanscrit as
closer to PIE than other I-E languages). This will tend to lead to
errors of the GIGO type in reconstructions higher than Proto-Altaic.

> My personal hunch so far is that Korean split away from Altaic the
> earliest and retains more archaicisms. Japanese, as we all know,
> is a highly innovative language and was probably affected by a
> native element. Overall though, I think the problems in Altaic
> are exaggerated. What is _really_ gnawing at the roots of the field
> is the lack of creative solutions combined with people who shouldn't
> be reconstructing the language. Starostin and his reconstruction
> of Altaic "three" word comes to mind as a great example of
> linguistic alchemy.

> This is *exactly* the kind of problem that I was referring to above.
> I've arrived at a better solution: Nostratic *t, *k and *p
> have softened in PreAltaic (*s, *x, *f), while the ejectives
> have become unvoiced stops and the voiced stops have remained
> unchanged. So far, the Altaic cognates used in Bomhard's etymologies
> are weak at best.

Agreed, but then in the absence of a sound study of intra-Altaic
phonemes, your schema is possibly no better (or no worse) than
Bomhard's or Starostin's.

> I don't think of any language as truely isolated, hidden away
> in the mountains somewhere. Like most languages, and especially
> for a language smack dab in the _center_ of Asia, Altaic must
> have had other neighbours that influenced it. These neighbours
> would have traded with the Altaic speaking peoples. Any
> terminology denoting innovative things would more likely come
> from the south. The fact that "sheep" looks similar in
> both Dravidian and Altaic, without other real cognates in other
> northern languages like Uralic or IndoEuropean (*ker- doesn't
> count in my books), makes it look like a loanword more than
> anything else.

I would agree. Even languages like Basque show significant borrowing
from neighbours. One wonders how much was borrowed from neighbours
which no longer exist (eg. Iberian in the case of Basque, Hunnic in
the case of Altaic (some I know see Hunnic as a form of Early Turkic,
others as Early Mongolian, and still others as an independent
language in its own right - another split to the Turkic-Mongol-Manchu
split). I suspect including Hunnic from the Chinese transcriptions
could have a significant effect upon our reconstructed proto-Altaic,
as Hunnic words seem to have been been borrowed into both Turkic and
(via Uighur) Mongol. Archaeologically too it makes sense if sheep is
a word that comes from outside the region some time after the split
of the Mesolithic Nostratic base culture.

Warm regards

John