Alexis on Wald on Linguistic classification

[Larry Trask]

----------------------------Original message----------------------------
Sasha Vovin writes:
 
[replying to Johanna Nichols on relatedness]
 
> This is quite a revolutionary definition of relatedness. I used to
> believe that relatedness is demonstrated by recurrent phonetic
> correspondences established on the basis of basic vocabulary and/or
> basic morphology, and I trust that all major families were done in
> this way without any appeal to "chance", as the very existence of
> this correspondences would rule out the "chance".
 
Well, I too would certainly like to believe that all recognized
language families have been arrived at in this way.  And some of them
certainly have been: IE, Algonquian, Austronesian, to name a few.
 
But others have not.  Two that spring to mind are Afro-Asiatic and
Niger-Congo.  To the best of my information, recurrent correspondences
in phonology and/or morphology have never been demonstrated for these
families, and no significant reconstruction is available for Proto-AA
or Proto-NC -- or, rather, none which has won any degree of general
acceptance.  Indeed, Bob Dixon has recently been complaining that the
evidence available to support the African families generally, and
Niger-Congo in particular, simply does not resemble the state of
affairs that Sasha describes.  Instead, it appears, the families are
set up on the basis of a few recurring grammatical characteristics,
characteristics which involve actual morphological material in the AA
case but only typological features in the NC case.
 
Dixon's main point is that, so far as he can discover, the evidence
for NC, and possibly for other African families, is not
distinguishable from the effects of millennia of diffusion across
language boundaries.
 
Are there any Africanists out there who can put Dixon's mind at rest?
 
Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK
 
larryt at cogs.susx.ac.uk