LINGUIST List 2.265

Sunday, 2 June 1991

Disc: Pseudo-oblique, Hist lx, Hyouston,

Editor for this issue: <>

Directory

1. , Pseudo-oblique objects
2. , Word stability in historical linguistics
3. "ELISE EMERSON MORSE-GAGNE", hyouston

My thanks to those who have sent me comments on the syntax of range-specifying
NPs such as "between 45 minutes to an hour". The grammatical problem that
these things pose is that they resemble PPs but behave like NPs.
Semantically, the prepositions name beginning and end points on a scale,
rather than a relation between an NP and a verb or situation. Right now, I am
inclined to think of them as headless post-modifying PPs. So (1) behaves as
if it had the syntax of (2):

 (1) Between 45 minutes and an hour elapsed.
 (2) A time between 45 minutes and an hour elapsed.

One could still analyze the subject in (3) differently, i.e. more like the
subject of (4):

 (3) Between 5 and 10 minutes elapsed.
 (4) Approximately 5 minutes elapsed.

In other words, one could still claim that 'between 5 and 10' fits into a kind
of 'measure slot' in the NP. But if you have to live with (1) anyway, then
maybe (3) could be treated as a headless postmodifier, too.

The headless postmodifier idea might also help to illuminate the nature of
double-preposition constructions:

 (5) Set the timer to between 45 minutes and an hour.
 (6) Remove debris from around the pipe.

I.e. "...to a time between 45 minutes and an hour" and "...from the space
around the pipe". Any comments on this line of thought would be appreciated.

				-Rick Wojcik (rwojcikatc.boeing.com)

Some weeks ago, while I was away, Bob Poser posted a query regarding
claims (esp. those of Aharon Dolgopolsky) about the stability of
words with certain meanings. Poser correctly
points out that Dolgopolsky's study is marred by the fact that
he did not publish his data, that his sample of languages, while
large, was unrepresentative (only Old World languages), and that
for the languages in the sample we have to take his word for the
fact that he correctly identified the words which have remained
unchanged since the oldest reconstructed stage and those which
have been replaced. I personally think the last objection
is not very important, precisely because the sample was skewed
in favor of languages Dolgopolsky knows. More generally, I do
not know of a better study of word stability, and I think that
the purpose for which this study was designed was such that it
was not an unreasonable thing to do. The purpose was to devise
a way of identifying groups of languages which it is reasonable
to assume are related. It was NOT intended to supplant the
conventional methods of comparative linguistics which are used
to (a) prove beyond a reasonable doubt the existence of a relationship
and (b) reconstruct a proto-language (or fragments of one).

As Poser points out, Dolgopolsky's partially ordered list of
the "stablest words" (including such items as the first and
second person as well as interrogative pronouns, negation,
the numerals 2 and 3, body parts such as heart and tongue, etc.)
has been referred to in recent accounts of the work on remote
relations. (To be sure, in the (in)famous Starostin quotation
about the word for 'hand', we are dealing with a word which
is NOT on the Dolgopolsky list.) However, it should be noted
(and I think has been emphasized in recent press accounts) that
neither the Nostratic nor the Sino-Caucasian hypothesis (nor
even the less believable Dene-Caucasian and some other theories)
are based on the stability argument. Whether correct or not,
these theories are based on claims of massive and nontrivial
sound correspondences and morphological relationships involving
large chunks of the morpheme stocks of the languages involved.
In fact, it is possible to poke holes in parts of the stability
theory by pointing out that, for example, the numerals 2 and 3
are NOT shared by the different Nostratic branches, as Dolgopolsky
himself often points out in other contexts.

Having said this, it would be very important to get more work
on the whole stability question. My own feeling (and it is
supported by various kinds of data, not the least the fact
about 2 and 3 in Nostratic) is that stability of words is
dependent on culture, and hence not universal in the standard
sense of the word. Thus, if a given people did not count at all,
then it would not have words (much less stable ones) for any
numerals, and so on. Hence, I would be skeptical of lists such
as Dolgopolsky as having a completely universal status, but used
judiciously such studies could prove useful. It certain seems
to be a common assumption of ALL comparative linguists that I
know of that you base your work on basic alias core vocabulary
and hence that there is such a thing as basic or core vocabulary
and this tends to be stable enough to allow meaningful comparison.
It would be nice to have both a more precise formulation of this
and to know the limits of what we are allowed to assume.

(I should add that Dolgopolsky himself in a later article noted
that first and second person pronouns become quite UNstable in
languages spoken by certain kinds of highly stratified societies
(such those of Europe, S., SE., and E. Asia, but he concludes
that such exceptions to his claims are easy to contain. I think
that the problem is much more difficult, and indeed in principle
insuperable.)

Norval Smith's comment that the original, Scottish place-name was pronounced
--I mean, IS pronounced--[hustn] suggests the fate of the word "coupon",
which has initial [ky] in many peoples' speech despite coming from French
with a [ku]. According to the OED that very word in fact was an early borrowing
(whoops) which has survived to the present in Scots English but was re-borrowed
more recently into the rest of the language. How do Scots say it?

[End Linguist List, Vol. 2, No. 0265]