LINGUIST List 2.92

Sunday, 24 Mar 1991

Disc: Vowels and Stress

Editor for this issue: <>


  1. , Unstressed vowels in English
  2. , Re: Unstressed Vowels
  3. , Re: Responses: closed syllables
  4. Richard Ogden, the suffix '-ate'

Message 1: Unstressed vowels in English

Date: Tue, 19 Mar 1991 23:46 MST
Subject: Unstressed vowels in English
 I'd like to thank everyone who answered my request for information
about languages with unstressed vowel systems. In another message I
will summarize the information I've received so far and ask a few more
questions that have been raised in my mind. In this one, however, I'd
like to respond to some comments on my apparently provocative words
about English.

 Some of the responses to my query support my suspicion that there is
a fair bit of confusion, misunderstanding, and "myth" (A.M. Ramer's
word) about what unstressed vowel neutralization (and reduction) might
be. I think that we do not have a clear picture of, or wide-spread
agreement about, what the issues are. I am clearly not the only
person who thinks so (cf. J. Coleman's remark: "'Reduced' is not a
well-defined concept in phonetics or phonology") and is concerned
about it (cf. the thermal level of some of the responses to my remarks
about English, which suggests that the nerve that was hit was already

 What often happens in such a situation is that people respond to
questions that were not asked, or sometimes they legitimize questions
irrelevant to the inquiry by answering them instead of challenging
them. The remarks which follow are intended to focus more clearly on
my query and the data and ask for participation in clarifying what the
issues are. 

 Here's a repeat of the first part of my query: "I am working on the
relationship between stressed vowels and the unstressed vowels they
neutralize to (assuming that there is a generative relationship
between the two). I am looking particularly for languages whose
stressed vowels neutralize to *more than one* unstressed vowel.
 Languages whose vowels are all schwa in unstressed syllables,
presumably like English in this respect, are thus of little or no
interest to this study. Languages like Catalan, however, whose seven
stressed vowels neutralize to three distinct unstressed vowels, are
exactly what I'm looking for."

 Many people brought to my attention 

 a) "reduced" vowels in English and in other languages, or
 b) the existence of *contrasts* between unstressed vowels in

though my query asked about *neutralized* (not "reduced") unstressed
vowels and referred explicitly to languages which have a *generative*
relationship between stressed and unstressed vowels. Thus most of the
examples that were offered me, such as accept/except, ribbon/ribbin',
Rosa's/roses, and tory/toro, are not relevant to my query--none of the
unstressed vowels here have stressed "counterparts." Thus there's no
neutralization apparent in these words, and maybe no reduction either,
unless by "reduction" is meant simply a sort of muffled quality of
unstressed vowels; even to say that this reduction is phonetic
centralization as compared to stressed vowels begs the questions of
what "as compared to" might mean, and how the stressed vowels
participate in the comparison if they have no unstressed counterparts.
 These responses indicate, at least, that there is confusion about the
terms "reduction" and "neutralization," and perhaps that some people
consider them synonymous.

 An example suggested by J. Coleman, title/titular, is also not
relevant to my query, because the vowel "counterparts" do not have a
stress difference. 

 An example like lobe/lobotomy might be relevant if it were really
clear that these words have a generative relationship. Someone
suggested that I learn about the variety of unstressed vowels in
English by reading Bollinger's work, but failed to note that Bollinger
has argued *against* a generative relationship between them and their
putative stressed "counterparts", proposing instead that we think of
stressed and unstressed vowels in English as composing separate
systems, rather than try to generate one set from the other. I am
inclined to agree with that, or at least not to want to fiddle
endlessly with English, because there are so many unresolved and
potentially misleading problems in the "derivational" morphology of
English, mostly resulting from the fact that many of the examples to
be brought to bear are either learn-ed or latinate/norman-french
(i.e., have a historical but not necessarily synchronic relationship
in English) or both. That is, it's not clear that there is really a
generative relationship between, say, lobe and lobotomy; and, in this
example at least, there's also the possibility of orthographic

 However, I do admit to infelicitous wording in my statement about
English; I regret it; I apologize for it. I said "Languages whose
vowels are all schwa in unstressed syllables, presumably like
English...." I am of course aware that English has more than one
unstressed vowel--I don't have to search the literature for obscure
dialects to get that information: in my dialect too the unstressed
vowel in "tory" and "obey" is not schwa. This is the statement that
jangled the sore nerve. I submit that people who found it hard to
believe that anyone could make this claim about English should have
trusted their gut reaction, i.e., they *shouldn't* have believed it:
 it was an elliptical statement that seemed clear to me in the context
in which I made it.

 What I intended to say (and what most of my readers interpreted my
words as saying, since the literal interpretation is so unlikely) was
this: "languages whose vowels all neutralize to schwa in unstressed
syllables, presumably like English...." *This* is a controversial
statement worthy of comment and objection. I thought that my
inclusion of the qualifying word "presumably" would signal my
acknowledgment of this, but I guess it didn't. 

 But the reason that I was moved to venture such a remark is this: in
pursuing this research I have gotten the impression that the idea (the
"myth") that English vowels all neutralize/reduce to schwa in
unstressed syllables is, though little-examined, *widely assumed*.
(It's reminiscent, to me at least, of the idea that Eskimo has 200
words for "snow," widely accepted because apparently plausible, but,
according to G. Pullum and others, not true). I didn't consciously
make this remark just to see if anyone would object to it, though
perhaps that was a subliminal motivation, the Imp of the Perverse at
work, maybe. However, it is notable that only a few people, of the
dozens who responded to my query, commented on my English remark at
all, and those few didn't offer any *neutralization* data (except for
Coleman, who suggested rEject/rejEct, albeit as if it were the same
kind of example as title/titular). The other respondents,
presumably (?!), either picked up on "presumably" or accepted my
remark as presented. What all this suggests to me is that my
suspicion, namely that there is no well-examined body of knowledge
about these phemomena in our understanding, i.e., that there is a
"myth" abroad here, has some basis.

 I don't want to wrangle about whether English falls into the category
of languages I'm looking for. (And, again, if you want to know why
I'm looking for them, ask me for my preliminary paper on the topic,
which I presented at the last LSA meeting, or look up the abstract in
the meeting handbook.) But I hope this discussion has raised our
awareness of how little we know about what happens to vowels in
unstressed syllables, and I thank those who sent me references to
articles written by people who've looked into the matter (although
most of these concern reduction or contrasts between unstressed vowels
in English, and not neutralization). 

 I continue to welcome input from as-yet-unheard-from people about
languages that do have unstressed vowel *neutralization*. 

 Watch this space for the summary of what I've learned about
unstressed vowel neutralization in other languages.

Christine Kamprath
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: Unstressed Vowels

Date: Wed, 20 Mar 91 09:09:03 PST
From: <>
Subject: Re: Unstressed Vowels
Michaeal Kac mentions that unstressed /o/ in Russian becomes /a/ when 
unstressed, and unstressed /e/ becomes /i/. In fact, the work on vowel
reduction in Slavic languages is quite old, going back to, I believe, to
the work of A. Potebnja (1835-1891) and Baudouin de Courtenay (who founded
modern phonological theory). Russian has five stressed vowels--i, e, a, o,
u--and three unstressed vowels--i, u, a--according to traditional analysis.
Controversy has developed around the phonemic analysis of the unstressed
vowels. Russian phonologists call reduction of /o/ to [a] akan'e, and 
reduction of /e/ and /ja/ to [i] ikan'e. (Michael left out reduction of
so-called /ja/, which typically corresponds to a stressed [a] preceded by
a palatalized consonant or the palatal glide.)

Here is a simplified account of what goes on in Russian vowel reduction:
all non-high vowels reduce to schwa in unstressed position, and vowels 
following palatalized consonants assimilate to a fronted position. Russians
tend to here reduced vowels as phonemic /a/ and /i/, respectively. There
are several complications--e.g. schwa lowers to /a/ in absolute initial and
pretonic environments. Moreover, Slavic languages and dialects differ
considerably in assimilative and dissimilative vowel harmonies.

Richard Rhodes (U.C. Berkeley) has produced a nice analysis of vowel reduction
in modern American English, and I suggest that anyone interested in this
subject contact him. One interesting feature of his work is the observation
that some English speakers have a kind of vowel harmony affecting unstressed
vowels, not unlike the Slavic cases.
 Rick Wojcik
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Re: Responses: closed syllables

Date: Thu, 21 Mar 91 01:18:07 EST
From: <jdbobaljATHENA.MIT.EDU>
Subject: Re: Responses: closed syllables
Bruce Hayes work on Yupik is curculating in 
manuscript form right now and will constitute a chapter (?)
in his forthcoming book.

There is also the following source:

Krauss, M ed (1985)
_Yupik Eskimo Prosodic Systems_
Alaska Native Language Center
Research Paper Number 7

The Center is in Fairbanks, Alaska
(PO BOX 900111, Fairbanks, Alaska, 99775-0120
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 4: the suffix '-ate'

Date: Thu, 21 Mar 91 9:49 GMT
From: Richard Ogden <>
Subject: the suffix '-ate'
Mike Hammond talks about 'words ending in the suffix -ate' and explains
that such words are exceptional because they have a long vowel but are
not stressed.
What is the linguistic status of '-ate'? It seems to me that it has none,
it's an orthographic form which might tell you something about the history
of English spelling. So it doesn't seem like a good idea to make statements
about the phonology of English on the basis of this sort of 'transcription'.
--- are the '-ate's in 'Latinate', 'pontificate' and 'certificate' the 
same phonological object? Certainly not in my speech; I have different 
phonetic possibilities for each one of them. So I should be careful in
my terminology.

Richard Ogden
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue