--- In
cybalist@yahoogroups.com, Harald Hammarstrom <haha2581@...>
wrote:
> > I am definitely with Jens here... Monovocalic theory cannot be
excluded on
> > typological grounds since some languages *can* (even if it is
only a
> > possibility) be analyzed like that and also there is the
standard example -
> > if all the Khoisan languages died out before linguists came
there, I bet 99%
> > of all linguists would swear that phonemic clics are absolutely
impossibile.
>
> Piotr mentioned something and I can elaborate a bit. Khoisan is
NOT a
> genetic unity in the sense of Indo-European. It is label to cover
the
> following bona-fide genetic units/isolates:
> 1) Hadza
> 2) Sandawe
> 3) Central Khoisan
> 4) South Khoisan
> 5) North Khoisan
> 6) Hoa
> 7) Kwadi
>
> Kwadi is extinct and known only from some unpublished notes.
Güldemann
> has tentatively aligned genetically with Central Khoisan.
Hoa (=Hòã) is also written as '#Hoan' (# = palatal click) or '=Hoan'.
I've just been looking at Starostin's lexicostatistical analysis of
Khoisan using the (a?) 100 word list. About 5 items (I haven't re-
counted) were unusable, either because they merged with another
meaning or because the meaning was not typically represented by a
single word. #Hoan and North Khoisan score 43% similarity. The
better documented South Khoisan group, the Taa group, scores 32%
with North Khoisan. The Central Khoisan comparisons with North and
South Khoisan score around 20% (3 by 2 comparisons, ranging from 15%
to 24%). From a North Khoisan perspective, this can be described as:
1) Central Khoisan split off from Proto-Southern African Khoisan
(PSAK).
2) Then South Khoisan split off.
3) Then #Hoan split off.
4) That left North Khoisan.
Starostin does compare the level of relatedness to Altaic rather
than to Indo-European.
The relationships bewtween Sandawe, Hadza and SAK seem to be equally
distant. I counted 11 matches between Sandawe and Hadza, and
Starostin summaries the match level between the three branches as 8%
to 12%. Austronesian gets that bad, and I think this is a better
score than Nostratic achieves.
I haven't discussed Kwadi - I presume the problem here is just that
we only have a limited knowledge of it.
Richard.