LINGUIST List 3.362

Fri 24 Apr 1992

Disc: Greenberg and Mass Comparison

Editor for this issue: <>


  1. , T

Message 1: T

Date: Sun, 19 Apr 92 23:12:42 EDT
From: <>
Subject: T

In response to some issues raised by Poser in his response to
my posting regarding the Greenberg classification:

(a) When I mentioned that Dolgopolsky had developed a more
sophisticated and more precise version of what Greenberg appears
to be doing, I was indeed referring to the article in Voprosy
Jazykoznanija (most of which has been translated into a kind
of English and published in the Markey/Shevoroshkin volume).
Dolgopolsky explicitly defines which sounds we are allowed to
count as "similar" for the purposes of mass comparison. He also
appears to insist that mass comparison be based on words with
identical dictionary meanings. Finally, he suggests a list of
some words which he regards (on the basis of an empirical if
rather impressionistic study) as the most likely candidates for
the sorts of words a language will tend to retain (with only
phonetic changes) rather than replace (via borrowing or neologism)
over long periods of time (e.g., 'heart' but not 'hand'). And
he urges that these words in particular be looked at in mass
comparison of the Greenberg sort. I might add that various
"grammatical" elements (such as CERTAIN personal and interrogative
pronouns) are high on the list, whereas (as I may have noted in
an earlier posting) there are cases where Greenberg classifies
a language without looking at ANY such morphemes (viz., Zuni).

(b) I guess I really asked for it when I pointed out that
my comparative knowledge of American Indian languages involved just
Uto-Aztecan. It is easy, as friend Poser has done, to list
lots of experts who have done better. But that, of
course, was not the point. The point is that very few people
have done deep comparative work on more than one (OK, two) American
Indian language family, though quite a few people know something
about more than one family (even me, I am afraid). The same
is, by and large, true of other recent proposals for large
(and old) language families, such as Nostratic and Sino-Caucasian.
Plenty of people know something about Altaic as well as Indo-
European as well as Uralic (or some such combination)--and
many more hold strong views about them (esp. about Altaic or
rather about the absence thereof). But very few people (if any)
have done deep and precise work in each of the proposed subfamilies
of Nostratic (or Sino-Caucasian).

(c) I do concede the main point about Papuan linguistics. It
is true, I now shamefacedly realize, that Greenberg's supporters
argue that his African classification has been accepted by
the run-of-the-mill Africanist, so we should accept his American
one, but that his opponents are by the same logic entitled to
counter by pointing that his Pacific classification is not
accepted by the run-of-the-mill Papuanist.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue