LINGUIST List 11.590

Thu Mar 16 2000

Disc: Underlying Schwa?

Editor for this issue: Karen Milligan <karenlinguistlist.org>

Directory

1. ewb2, Re: 11.584, Disc: Underlying Schwa?
2. Todd O'Bryan, Re: 11.584, Disc: Underlying Schwa?
3. A.F. GUPTA, Re: 11.584, Disc: Underlying Schwa?
4. Jorge Guitart, shwas underlying shwas--once more with feeling

Message 1: Re: 11.584, Disc: Underlying Schwa?

On the question of one or two reduced vowels in English:
my North American English (New England variety) certainly
has two. 
Hit 'im means "hit him", although the vowel in the second
syllable is not identical with that in "him".
Hit 'em means "hit them", although the vowel in the second
syllable is not identical with that in "them".
The two utterances are distinct: the first has [I] and
the second [].

Wayles Browne, Assoc. Prof. of Linguistics
Department of Linguistics
Morrill Hall 321, Cornell University
Ithaca, New York 14853, U.S.A.

tel. 607-255-0712 (o), 607-273-3009 (h)
fax 607-255-2044 (write FOR W. BROWNE)
e-mail ewb2cornell.edu

Message 2: Re: 11.584, Disc: Underlying Schwa?

Responding to Jakob Dempsey (and adding to Max Wheeler):

I'm an American English Speaker(tm), and I definitely have the two reduced
vowels,  (schwa) and 1 (barred-i), in my pronunciation. Jakob is correct
that there's never a minimal pair based on this distinction, but I really
have to concentrate to produce the wrong vowel in a given environment.
Strangely enough, most of Max's judgments seem right to me. Just so you
know, I grew up in Louisville, Kentucky, so my native accent is nothing like
Max's in most respects: I have rhotics all over the place, and have strange
quirks of Southern English like losing the distinction between I (small-cap
"i") and E (epsilon) before nasals. "pen" and "pin", "hem" and "him" are
both pairs of homophones for me.

I thought everybody had the distinction between these two, and was shocked
that some people didn't. I was TA-ing an intro class where the professor
didn't have the distinction and only taught  as the reduced vowel. You have
no idea how many I (small-cap "i"'s) showed up in the homeworks in reduced
positions. People had trouble hearing certain sounds as schwa's even after
being told it was the only reduced vowel.

Responding to Theo Venneman, partially reproduced here:

> Another part of the confusion results from the belief that the feeling of
> relatedness of words has to be expressed by assigning those words identical
> (parts of) so-called underlying representations. This feeling of
> relatedness of words, which varies from one speaker to another, results
> from the words' phonological similarity and similarity of meaning.

I think you have discounted another perfectly valid hypothesis, and the one
which drives those of us who believe that speakers' judgments about
relatedness (at least those that are universal or nearly so throughout a
speech community) SHOULD be reflected in the grammar. You say that a feeling
of relatedness results from the words' phonological and semantic similarity.
Is it not also valid to suggest that the words' phonological and semantic
similarity results from their relatedness? Thus, "syllable" and "syllabic"
are in some way cut from the same cloth. Representing this cloth as an
underlying representation allows the (I would hope) nearly universal feeling
that "syllable" and "syllabic" are related to one another in a way that
"syllabus" is not related to either to be somehow encoded in a description
of what speakers know.

Are only words which differ in some vowel quality subject to this encoding
as underlyingly distinct, or are words with inflectional affixes similarly
felt to be related because they happen to sound alike and mean strikingly
similar things (e.g., dream, dreamt, dreaming, dreams)? In other words,
would you not want to encode the underlying similarity of these words
somewhere in the grammar, or do you leave open the possibility that there is
some proficient speaker of English who goes along blithely using these forms
correctly as completely different lexical items with completely unrelated
meanings?

Todd O'Bryan
UC-San Diego
obryanling.ucsd.edu

Message 3: Re: 11.584, Disc: Underlying Schwa?

 maxwcogs.susx.ac.uk (Max Wheeler) pointed out that many speakers of 
English quite naturally mark some distinctions in unstressed 
syllables which other speakers of English don't mark. In my variety 
of British English (Northern) I have most of the I/ distinctions Max 
has. Like many northerners I have a few more. These three are 
different:

rosers (i made this word up), roses, Rosy's
taxers, tax is, taxis
z, Iz, iz

I also have a full vowel in the first syllables of EXAM (CV3), 
CONFESS (near CV6, same as the COT vowel). But as Max said, speakers 
of one dialect can't assume that just because they make, or don't 
make, a distinction, other people don't or do!

Anthea
 * * * * * * * * * * * * * *
Anthea Fraser GUPTA : http://www.leeds.ac.uk/english/$staff/afg
School of English
University of Leeds
LEEDS LS2 9JT
UK
 * * * * * * * * * * * *

If every shwa is underlain by a shwa, then there is no need to say that
there is a process of vowel reduction, shwa formation, or whatever you
want to call it: there is no 'V becoming a shwa' since V is already a
shwa. We can do that with every single segment of every single lexical
element and turn phonology into tautology: /X/ is [X] in every case. There
is no case in which /X/ is [X] and [Y].

A friend from Galicia just sent me a paper on rhotization of /s/ in
Galician. /s/ is pronounced as a flap [represented here as r], e.g. in
front of [d] so that los dos ('the two of them') is pronunced lor dos [This also happens in
Castilian Spanish by the way]. 

But how could /s/ possibly underlie [r]?
The UR of [lor] would have to be /lor/ (as the UR of cactus would have to
be kakts where  symbolizes shwa and cactus and cactusia do not share
kakVtus where V is whatever) So then, you simply memorize that
the word los (which is something abstract with no phonological content) is
pronounced los in context A , lor in context B etc. There is NO principle
(rule or contstraint,take your pick) specifying that /s/ is pronounced [r]
in Galician. 

I prefer the (to me much more imaginative) non-memory solution: underlying
both [los] and [lor] is /los/. You 'know' that [lor] is /los/ because you
know that the principle '/s/ must be pronounced as [r] in context X' has
applied. 

Jorge Guitart