Bantu & losing tones, was: niKòmbá changes

On Mon, 29 Jan 2001 14:31:46 -0500, David Stokes <dstokes@...>
wrote:

[...]
>        I know almost nothing about the development of Swahili, but it lost
its
>tones
>while most other Bantu languages have them. Does anyone on the list know
>the
>development it went through ?
Try the CBOLD site:

http://www.linguistics.berkeley.edu/CBOLD/

I think comparing the proto-forms in their database
with the words from extant languages may give you
a few insights.

proto:
http://www.linguistics.berkeley.edu/cgi-bin/CBOLD/jgspltscrn.cgi?simpledbq=b
lr2.txt

Swahili (a bit tricky - it's part of Tanzanian Survey):
http://www.linguistics.berkeley.edu/cgi-bin/CBOLD/jgspltscrn.cgi?simpledbq=G
LOSSES.TXT

As for transforming tonal systems into smth different -
a couple of ideas not related to Bantu:

a) shift to stress position rules; e. g. rightmost combination
'high mora + low mora' in a word draws the stress onto its low
mora (or any other weird rule).

b) differentiate changes for consonants belonging/adjacent to
high vs. low moras (probably, more 'weakening' for low).

c) reduce unstressed vowels (no matter where you put the stress)
to different degrees depending on their tone (probably, more
reduction for low).

d) diphthongize vowels carrying 'uneven' tones.

Usually 'high' implies 'tense' and vice versa.


Basilius