Even the monophyletic status of South
African Khoisan without Hadza and Sandawe (both in Tanzania) is widely
questioned. Khoisan may be more realistically treated as a Sprachbund
involving perhaps three different families whose relationship to one another is
unclear. Those who assign Hadza and Sandawe to Khoisan classify them as
coordinate sister branches of the whole South Africal phylum. The main reason
for adding them to Khoisan is that they have clicks and are not otherwise
classifiable (not even as related to each other). However, the Nguni branch of
Bantu is also rich in clicks, and they are found in one South Cushitic language
(Dahalo, in Kenya), so the mere existence of click phonemes cannot be
considered a family diagnostic. The phonotactic properties of
clicks in Hadza and Sandawe are different from those found in
Khoisan of Nguni: Hadza and Sandawe permit them in non-initial positions. It's
likely, given the modern distribution of click languages, that the enclaves in
East Africa represent a residue of what was once a wide and continuous range of
occurrence, with numerous "click families".
Piotr
----- Original Message -----
Sent: Wednesday, October 09, 2002 1:15 PM
Subject: Re: [tied] Let dogs have their day too
You have any more info on Sandawe being an isolate Piotr? I
thought it was as safely Khoisan as any Khoisan lang. is, although being far off
to the north. It is true that they have the clicks
though!?