Trask on Dixon on African lgs--And extremism generally

[bwald]

----------------------------Original message----------------------------
I read the exchange challenging the "Niger-Congo family" theory with some
surprise.  I can't find all the pieces of it now because they are under all
kinds of different Re: headers, but I remember the gist of it.  Here's a
typical quote, purporting to report Dixon's views:
 
>Larry Trask says about the language families of Africa (posited by
>Greenberg and
>for the most part widely or even universally accepted by competent
>observers): "...
>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."
 
To avoid possible terminiological confusion, first note that Niger-Congo
and Niger-Kordofanian refer to the same proposed family, in the same way
that Indo-European and Indo-Hittite do/have.  When Kordofanian first came
to the attention of scholars, like Hittite and IE, it was thought to be so
remote that it had to be grouped apart from a grouping of all the other NC
languages.  NB: the SUBGROUPING problem.  One reason could be the suspected
link in Kordofanian between NC and Nilo-Saharan.  As I mentioned
previously, the "link" was later reclassified as Nilo-Saharan, and removed
from  "NC" Kordofanian languages.  Since then, most scholars most
immediately concerned have reverted to the NC label, with Kordofanian and
Mande problematic for which "split away" first.
 
With regard to the basis for NC, it is absolutely not true that it was
based on "typological" characteristics.  Indeed, when Greenberg first
proposed NC, many Banutists resisted the idea that many languages which are
now classified as "Bantoid" (along with "Bantu") but have lost the complex
prefixal noun classification system and various other inflectional systems
(these Cameroonian and Nigerian languages were previously called
"semi-Bantu", see Greenberg on that being like calling Icelandic
"semi-English" or something of that sort) were closely related to Bantu,
indeed at all related, rather than "mixed" with Bantu (massive borrowing, I
guess).
 
Recognition of NC was and remains based on what looks like numerous
cognates in all branches, so many that mistakes in sub-classification were
made on the basis of shared cognates which seemed to be innovative.  Many
of them turned up later in excluded branches, changing people's minds about
previous sub-classifications.  Kay Williamson's overview article in the
Bendor-Samuel book "The Niger-Congo Languages" Lanham: NY 1989 discusses
some of that, and the book as a whole is still the most comprehensive
general discussion of the family.
 
Meanwhile, Mukarovsky in Vienna has done and published extensive lexical
cognate hunting and gathering for the various branches of NC.  It is true
that reconstruction of the phonology of NC has not yet met the standards it
MUST in order to satisfy the proper demands of establishing the family  --
largely due to Scott's maxim "too much data, not enough scholars" -- but
the results so far look promising.  Also, despite what one might expect,
even on the basis of individual branches and even the phonologicasl
variability in Bantu, solving problems of cognate-dom do not look any more
difficult than for IE or other such groups.  It is worth noting that
Mukarovsky is so impressed by the mass of promising cognates in most
branches, that he is a severe critic of inclusion of Mande and Northern
Atlantic (but not Southern Atlantic) in NC.  That is, he was not able to
recognize sufficient candidates for cognate-dom in those branches to
justify their inclusion in NC.  (He did not survey Kordofanian at that
time.)  And he has since published against the notion -- though I don't
know what his latest views are.
 
NC scholars, Mande scholars, Atlantic scholars are NOT disturbed.  There is
plenty of work to do in reconstructing those families, whether they are
branches of NC or not.  The general attitude is that eventually the truth
will be discovered one way or the other, and much will be learned during
that trip.  Meanwhile, the work remains to be done and continues to be
done.
 
The typological similarities of various branches of NC is gravy, and I
think someone else already said that it is NOT true that grammatical
morphemes are only typologically similar in the various branches.  They are
similar in shape with similar meanings and promising for cognate-dom.  The
*ba- class prefix for human plurals is easily and regularly seen throughout
the group, etc etc.
 
Given my interest in historical syntax, I have found that consideration of
other NC languages adds much to my understanding of Bantu syntax, its
origin and evolution, EVEN IF NC does not hold up as a genetic group.
After all, no one in their right mind (but who is?) would claim that the
grammatical patterning of a language MUST have the same origin as its
lexical material.  Anyway, I'm not too worried about that because the
detail is convincing of genetic relationship, including the forms, not just
the uses of the forms in grammatical processes.
 
It is revealing to me to see how the (in my view, expectable) shortcomings
of the current status of NC are used by some scholars who seem not to be
interested in NC but in somehow using the status of NC to argue with and
damage each other.
 
As for Greenberg, because it is obvious that some want to discredit NC in
order to discredit what others considered his most successful
classification, all NC-ists appreciate what he did for us.  Where he was
right he was right, and where he was wrong we found out he was wrong.  But
he was right enough to be a great help.  (And it helped that what came
before really sucked.)  That does not mean I or any other NC-ist accept
what he says just because he says it, without knowing anything else about
it.  Our own experience is sufficient to prevent us from doing that.  As
for Merritt, in my opinion, what's good in Greenberg is good in Merritt and
what's bad in Greenberg is bad in Merritt.  He was that good a student of
Greenberg's.  Enough said for the moment.