Language and Anthropology in the Americas

[manaster at umich.edu]

----------------------------Original message----------------------------
My own view (based on I think not insignificant
1st-hand experience with American languages as well
as Old World ones) is that it is impossible to
overstate the extent to which American Indian lx
lags behind the Old World linguistics (always
with the exception of work by Bloomfield, Sapir,
and sometimes Whorf).  There are hundreds at least
linguists working on IE (and God knows there are
many unsolved problems there for them to work on) alone,
while in the Americas, esp. when it comes to comparative
work and even more so classificatory work, we are talking
of dozens or hundreds of languages per linguist.
Take Uto-Aztecan.  There are two people alive that
I know of who have worked on its possible external
connections to other language families, aside from
Greenberg, for a total of I think three published
papers over a period of decades, within UA, more
people work on Nahuatl than all the other languages
put together, and most of this is synchronic,
literary, or detail work.  There are very few
people doing comparative work across the family
as a whole.  In addition, despite all teh descriptive
work, we still do not possess adequate data for most
languages.  For many body part terms even, I canot
find ut what the word is in many if not most UA
languages.  The amount of info we have on Sanskrit,
Latin, Greek, and perhaps even such poorly attested
languages as Hittite vastly exceeds what we know
of almost any Amerindian language, so there is
a very severe limit on what we have to compare.
It was only a few years ago, with UA linguistics
entering its second century, that I found out
from a very obscure source the Tubatulabal word
for 'tear', for example.  Nor are teh gaps just
lexical.  We lack descriptions of the morphophonemics
of most languages, and bear in mind that morphophonemics
is by its nature usually a repository of information
about the past stages of teh phonology.
 
The Austerlitz-Nichols position, which Miguel seems
to be adopting, seems to be (this is expliclity
stated by Austerlitz) based on the assumption that
the opposite is teh case, that is, that American
Indian languages HAVE been investigated as fully
as Old World ones and that if there were easy
linguistic relationships to be found (say comparable
to IE or Uralic in depth) they would have been found.
As I say, Austerlitz says this explicitly, but this
is just absolutely incorrect.  I do not know
(aside from Haida-Nadene, Pakawan, and Pakawa-
Karankawan) just how many such relationships
have in fact been staring us in the face without
being found and acknowledged, but it is clear
that nothing even remotely approaching the amount
of work done on Old World problems has been
done here.  We are in the Americas at the stage
where Strahlenberg (whom Jakobson called the
father of comparative linguistics) entered
the arena of classifying the non-IE languages
of Northeastern Eurasia in the early 18th century.
 
Of course, we must always make that exception for
Sapir, but there again after 1920 or so
Sapir stopped providing adequate documentation for
his classificatory proposals, and so all his
proposals from that time on are in need of
reexamination from scratch.
 
So anybody who assumes that the universally accepted
language families of the New World can be compared to
those of the old and that consequently any lumping
to be done in the New World would be the moral
equivalent of Altaic or Nostratic is just ignoring
the whole history of the field.  Most recognized
New World families are the size of Romance or Slavic
or less.
 
AMR
 
On Sat, 9 May 1998, Miguel Carrasquer Vidal wrote:
 
[snip]
>
> I *have* read "Language in the Americas", and for whatever it's worth
> (I have no first hand experience with any American Indian language),
> the impression I got was that Greenberg had plainly failed to make a
> convincing case for Amerind.  That doesn't mean that Amerind *is*
> invalid, of course.  But considering that the time-depth of
> Proto-Afro-Asiatic might well be 10-12,000 years, I think
> (impressionistically!) that the odds that Amerind is a genetic unit
> at a comparable time-depth are very small indeed.  Whether the real
> time-depth is 20 or 50,000 years is anybody's guess, and depends
> largely on work that has yet to be done (how much more than 3 and how
> much less than 58+17+118 is the real number of "medium-range" genetic
> units in the Americas?).
>