LINGUIST List 5.1479

Mon 19 Dec 1994

Disc: Rate of Loss for "Basic" Vocabulary

Editor for this issue: <>


  1. , Rate of Loss for "Basic" Vocabulary

Message 1: Rate of Loss for "Basic" Vocabulary

Date: Wed, 14 Dec 94 21:45:55 ESRate of Loss for "Basic" Vocabulary
From: <>
Subject: Rate of Loss for "Basic" Vocabulary

Some months ago there was an extended discussion of the
proposal made in the fifties by Swadesh that there are
certain lists of meanings which have the property that
the words expressing them are lost at a fairly constant
rate per millennium in ALL languages, in particular, a
certain 100-meaning list where the rate is around 86%.
While noting that Bergsland and Vogt have adduced
examples where the rate seems to be LOWER, I pointed
out that I had not found any examples where it was
clearly higher. Jacques Guy referred to the Eastern
Greenlandic data discussed by Bergsland and Vogt in
Current Anthropology in 1962, but, having just carefully
read this paper, I see that they state that, while they
surmise that the rate here was higher, they cannot
calculate what is was(presuumably because there are no
sources for Greenlandic that are old enough). Moreover,
the purely conjectural rate of 72% which they mention is
still low enough for the purposes of the original discussion
we were having about this, namely, to assure that in a family
with enough ramification there would be a substantial portion
of the original vocabulary reflected in at least two descendants
after far more than 10,000 years (provided each language
was independely losing 28% of the original vocabulary per
millennium). I don't have the numbers handy, but Jacques,
who is better at this anyway, can probably whip some up.

But I should also add that Bergsland nd Vogt's data
are also consistent with many other assumptions, notably,
that the rate of retention in E. Greenlandic was 86%,
but that it was substanitally LOWER in some other Eskimo
languages (because all they have is the result that E.
Greenlandic is as different lexically from W. Greenlandic
as the two together are from Yupik, which makes no sense
given the family tree if the rates are the same eveywhere).
But this could be due to slower vocabulary loss in Yupik
(or even in the common ancestror of E. and W. Greenlandic,
or both) just as well as to faster loss in E. Greenlandic.
So there is no justification that I can find for Bergsland
and Vogt's or Guy's assumption that the "prolem" lies
in a faster rate of loss in E. Greenlandic.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue