LINGUIST List 11.584

Thu Mar 16 2000

Disc: Underlying Schwa?

Editor for this issue: Karen Milligan <>


  1. Max Wheeler, Re: 11.553, Disc: Underlying Schwa?
  2. Theo Vennemann, Re: 11.572, Disc: Underlying Schwa?

Message 1: Re: 11.553, Disc: Underlying Schwa?

Date: Wed, 15 Mar 2000 16:59:03 +0000
From: Max Wheeler <>
Subject: Re: 11.553, Disc: Underlying Schwa?

- Begin original message --
> a) Two different kinds of schwa ?
> In my very native style of American English I fail to find this
> division into two different kinds of unstressed vowels Jorge
> mentions. For example, in "pallet" and "ballot" I pronounce everything
> the same except for the initial consonant (I would transcribe as /
> pal't, bal't /. Is it a mere coincidence that, as far as I know,
> every case of this supposed high-front unstressed vowel corresponds to
> a syllable where the vowel is SPELLED with " i " or " e " ? In other
> words, the existence of these two kinds of unstressed vowels may be no
> more than an affectation by the highly-literate speakers who speak and
> teach it. I feel that the major task of phonology, dealing with a
> "low-level" of the speech phenomenon, is concerned with unconscious
> behavior, not highly self-monitored, self-reported behavior. ("OK,
> try saying 'rabbit' REAL SLOW now!")
> In any case, since the minimal pair, by which we define
> vowel-phonemes, is a matter of stressed syllables, is there any great
> significance is this slight, unstressed variation which some speakers
> claim to have? Would it even ever disqualify someone from the title of
> "native speaker"? Just as a matter of curiosity I would be interested
> to hear if this sort of distinction in unstressed vowels has been
> measured experimentally (I mean with acoustic equipment) in large
> samples of unmonitored speech (preferably U.S. English). As far as I
> know, it's simply what the dictionaries tell us.
> - -Jakob Dempsey
> Yuan-ze University, Taiwan
> ---------------------------------------------------------------------------
> LINGUIST List: Vol-11-553

- End original message --

The contrast (unstressed I vs ) may not be common in contemporary American
English but it seems extreme to claim that it's artificial, given that in other
dialects this distinction is clearly maintained, before a wide range of
consonants (though not all). In my British RP (non-rhotic) I have contrasts like
the following, not all of which are reflected in spelling (though most are).

	Philip dollop
	Carib Arab
	serif seraph
	(restive rest of)
	passim possum
	zenith penn'o'th (=pennyworth)
	pallet ballot
	(palate is variable I ~ )
	batted battered
	valid ballad
	penis Venus
	catches catchers
	chicken thicken
	tendril spandrel
	frolic bollock

I do not think I have a contrast before alveo-palatals, so the following all have
[I] and there are no counterparts with []:

More generally I don't expect words which are perfect rhymes for me (such as sofa
- loafer) to be rhymes in other dialects, nor do I suggest that those who make
such distinctions do so 'artificially' or on the basis of spelling, which is not
to say there may not be SOME such people.

By the way, for me, hermit and permit (N) are perfect rhymes (with [I]), so I
never understood Chomsky & Halle's argument based on the distinction between the
vowels in the second syllable. 

Max Wheeler

Max W. Wheeler
School of Cognitive & Computing Sciences
University of Sussex

Tel: +44 (0)1273 678975 Fax: +44 (0)1273 671320 Email:
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 11.572, Disc: Underlying Schwa?

Date: Thu, 16 Mar 2000 11:13:24 +0100
From: Theo Vennemann <>
Subject: Re: 11.572, Disc: Underlying Schwa?


Response to Jorge Guitart and Larry <>

Vowels in reduced syllables cannot be anything but shwas, because reduced
syllables by definition do not permit any individual vowel color in their
nuclei. This is so independently of whether there happen to exist words
felt to be related to the words under consideration and having a full vowel
in the corresponding position. Certainly the invention of a word such as
"cactusia" based on "cactus" has no effect on the phonological structure of
"cactus"; it merely indicates that the inventor knows how to spell the word
"cactus". But the same is true of an established word such as "syllabic"
alongside "syllable" or "syllabus". That "syllabic" has a full vowel in its
second syllable is a phonological property of this word, whose second
syllable is not reduced and thus permits (indeed requires) a full vowel,
but not of words felt to be related that have a reduced second syllable.
The second (and third) syllables of "syllable" and "syllabus" are and
remain reduced under all circumstances. Therefore, if you use shwa-ness as
an exponent of syllable reduction, this is the same as saying The vowels of
the second (and third) syllables of "syllable" and "syllabus" are and
remain shwas under all circumstances.

In short, and accepting the question as it was intended: All English shwas
are underlying shwas.

This applies to word phonology. In sentence phonology shwas may derive from
full vowels by realizational processes.

Part of the confusion behind the original question results from the
mistaken assumption that shwa is a vowel phoneme of the English language.
It is not. English has a prosody of syllable reduction: A reduced syllable
of a word is one whose nuclear position permits no individual vowel color.
This neither increases nor diminishes the number of English vowel phonemes.

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. Thus,
many speakers of English feel the words "syllable" and "syllabic" to be
related. This feeling need not be expressed in a description of English by
making the two words appear even more phonologically similar than they are,
namely by assigning them the same full second vowel. It need not (and
should not) be expressed in a description of English at all, because it is
an individual property of speakers of English to feel this relatedness and
not a property of English. By representing the phonological and semantic
properties of the words "syllable" and "syllabic" independently of each
other the basis will be laid for an explanation of how some speakers feel
about these words.

With kind regards,
Theo Vennemann.
6 March 2000
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue