Amerind (was: Re: Verb order in Montreiano)

Andreas wrote:

 > Hm, could anyone tell me some of this "Amerind Hypothesis"? All I know of it
 > is that I was once attacked for adhering to it, while my only mistake was
 > following the example of a history book from school in using "Amerind
 > peoples" as a catch-all term to include Red Indians, Eskimoes and Aleutians"

"Amerind" is used differently in History and Anthropology than in
Linguistics. Whereas in most fields, "Amerind" refers to all the native
people of the Americas, in Linguistics the term refers to a genetic groups
proposed by Joseph Greenberg at the end of the 80s. (Brilliant typologist,
weak historical linguist) It was highly controversial and rejected by
virtually the entire linguistics community. (I don't know of any
specialists in American Indian languages who believe in it.) For some
linguists (including my advisor) this is a very loaded term that should not
be used to refer to languages of the Americas. I personally would not
recommend refering to a language as Amerind unless you mean to imply that
you buy Greenberg's proposal.

The hypothesis is that all the languages of the Americas except those in
the Eskimo-Aleut and Na-Dene families are decended from a common
proto-language: "proto-Amerind". He did this by compiling lists of words
from dictionaries of lots of these languages, and comparing them to see how
similar they looked. He drew his conclusion on these 200-odd comparisons.
He never tried to find regular correspondances in the words (a basic
requirement in historical work), and the meanings of the words varied wildly.

To choose a random example, his examples for CHILD1 (out of 4 sets of words
for CHILD) have the meanings: copulate, son, be born, bear, girl, small,
tender, child, boy, and young (of animals). CHILD3 contains such meanings
as: ferment, grow, beget, be born (of a plant), grandchild, nephew,
pregnant, children (yes, he had to use a plural form to make the words
similar), sprout, and daughter, in addition to the examples just given for
CHILD1.

 From the same set of words, he lists words that vary phonologically like:
nin, ani, ni, ineu, anax, naijihi, añu, and pauna. And for CHILD3: pui,
aibo, fe, pom, pu, Baki, bakwa, poj, par, p'amp'an, apuhudn, pini, abje,
(n)(')kwa(h)(n). As you can see, the tendancy is for there to be similar
consonants in each word (usually only one), while the vowels vary wildly,
free to occur before or after the stable consonant.

It is certainly possible for such forms to be related, but you have to find
some kind of systematic regularity to prove that. Greenberg never even
tried to find any. His comments on the proto-grammar are equally weak:
there is no evidence of a common case system, so Amerind probably did not
have one.

Something I found interesting is the fact that he never justified his
decision to exclude Eskimo-Aleut and Na-Dene.


Marcus Smith

"Sit down before fact as a little child,
be prepared to give up every preconceived notion,
follow humbly wherever and to whatsoever abysses Nature leads,
or you shall learn nothing."
                -- Thomas Huxley