LINGUIST List 2.83

Wednesday, 20 Mar 1991

Disc: Language Families, Vowels & Stress

Editor for this issue: <>

Directory

1. Toby Paff, Nostratic
2. Scott Delancey, Re: Language Families
3. Mark Durie, RE>Language Families
4. "Bruce E. Nevin", Re: Curious Stress Patterns
5. Rachelle Waksler, Re: Stress and Huasteco
6. Mike Hammond, Sinhalese Stress
7. 71040000, closed syllables
8. "Michael Kac", Re: Unstressed Vowels

Some discussion of the extended language family hypothesis can be found
in an issue of Science that appeared last summer (sorry I lack the exact
reference here ... any serials index can get it for you). The discussion
is more interesting than the usual in that it also takes a look or two
at the supposed 'biological' evidence as well, which, though I have no
way of knowing directly, is presumably at least as shakey as the linguistic
evidence. The discussion stems from the conference that was held to deal in
part with some of Greenberg's work.

Toby Paff
(tobypaffpucc)

While I haven't read any of the journalistic coverage of the issues
referred to in Dan Everett's note, I must say that as a contribution
to the discussion so far in this forum it seems a bit intemperate. We
should all heed Manaster-Ramer's caveat about not lumping all lumping
proposals together; whatever the merits and deficiencies of the work
of linguists like Starostin and Benedict, it is unfair to tar them
with the same brush as Greenberg, whose work is (and I am attempting
to be maximally neutral and diplomatic in my phrasing here) methodologically
innovative in ways that Starostin's certainly is not.
 It certainly is not true that no "mainstream historical linguist"
has ever engaged in work at this time depth, unless one automatically
takes any expression of interest in deep reconstruction as disqualifying
anyone from membership in th mainstream. Semitic-Indo-European connections,
in particular, have been explored for generations, by linguists such as
Carleton Hodge, who is as "mainstream" a historical linguist as I can
imagine. And I have been told (perhaps someone more knowledgeable can
comment) that a reasonable estimate for the time depth of Niger-Khordofanian
is on the order of ten millenia or so, which is the same order that would
be involved in Austro-Tai or, on Greenberg's hypothesis, in Amerind. (In
fact Greenberg is most likely wrong about the time depth of human 
occupation of the Americas, but surely no one can consider the Clovis
time depth to be a crank or fringe notion).

Scott DeLancey
University of Oregon

A note for Daniel Everett: since you are only familiar with the South American
work of Greenberg, how can you call the Soviet work on Asian families
unscientific drivel? Can you read Russian?

Goldsmith (1990) _Autosegmental & Metrical Phonology_ talks about this
on pp. 115-116, and cites Southeastern Tepehuan (Uto-Aztecan) as
reported in Elizabeth Willett (1982) Reduplication and accent in southeastern
Tepehuan, IJAL 48:168-84. CV syllables are the lightest, CVV syllables
are intermediate in weight, and CVC syllables are heaviest. He also
discusses a 3-way distinction whose best-known case "may be" Estonian.

Hope this is helpful.

 Bruce

Subject: Re: Curious stress patterns
Huasteco has a pattern that syllables with long Vs are heavier
than closed syllables. (Lahiri and Koreman 1988, a WCCFL paper)
I don't know whether there are any rules that will show whether
open short V syllables are lighter than closed syllables, but
it's a place to start.

With respect to Alexis Manaster-Ramer's recent posting in "Linguist"
about Sinhalese stress.

There are other languages where there is a ranking of syllable weights
beyond a simple distinction of light vs. heavy. I actually have a
paper in NLLT in '86 where I include a discussion of Klamath which
distinguishes light from closed from long syllables. You could even
argue that English nouns distinguish closed syllables from sylables
containing a long vowel. A final open light syllable is stressless in
English, e.g. America. A final closed syllable may be stressed or
stressless: helix vs. narthex. But a final long vowel must be
stressed: anecdote. (Words ending in the suffix -ate are exceptional:
certificate.)

mike hammond

The Yup'ik language has several dialects which make a distinction for foot
building where CVC will take stress over a CV syllable, but a regular foot
will be built over a sequence CVC CVC, or CVC CVV. Steven Jacobson is the
author of the Yup'ik dictionary, and there's a grammar, which might be by
him also, I can't remember. Bruce Hayes has done some work with it, but
I'm not sure that it's published.

A thought prompted by Coleman's comment that 'reduced' is not a well-
defined concept in phonetics or phonology: In one sense it is, at least
in the traditional way of talking about two languages I know a bit about,
namely English and Russian. In both languages, you get alternations in
which the stress changes (the English *telegraph-telegraphy* example is
a textbook case in point) and where segmental features of the stressed
vowel are different from those of the unstressed one. My earlier comment
about the 'Rosa's Roses' case is thus, in retrospect, beside the point:
there's an issue there as to whether the two unstressed vowels are the
same but that, I now realize, doesn't have anything to do with reduced
vowels in the sense just cited.

In Russian, at least according to the textbooks (and I'm told by those
who know more about it than I do that it's Muscovite Russian), underlying
/o/ becomes /a/ when unstressed and underlying /e/ becomes /i/.

Mea culpa.

MK