19

The first syllable in "about" (ə'baʊt) is schwa, so is the second one in the "salad" ('sæləd), but iv'e never heard them pronounced the same way.

in salad it sounds more like the i in "trick". mountain ('maʊntən) also has a schwa at the end, but to me it doesn't sound like about, much more like the sound in trick.

So is it a consistent sound?

  • Maybe this question fits better on English Language Learners or English Language & Usage – jknappen - Reinstate Monica May 10 at 9:20
  • 6
    @jknappen Stack sites overlap. This question is not offtopic here, and OP probably wants a linguist's POV than a writer or teacher one. Choice of stack to post a question changes the kind of specialist that will provide the answer. – Mindwin May 10 at 18:04
  • @Mindwin: If you really want to discuss whether a question is off-topic or not, this belongs to linguistics.meta.stackexchange.com – jknappen - Reinstate Monica May 10 at 18:15
  • 1
    @jknappen: Well, you started it! In my opinion this question is much better suited to Linguistics than either of the sites you mention. You tend to get decent answers here :-) – TonyK May 10 at 23:53
  • The Schwa was here. – Mateen Ulhaq May 12 at 2:35
28

No, it is not. The English sound transcribed as "schwa" (ə) is known to be quite variable. There are a number of things that affect how it sounds.

In American accents, it's common for a "schwa" to be more similar to an an [ɪ] sound (the "i" in "trick") when it's in the middle of a word. For many American English speakers, there is no consistently perceptible contrast between /ə/ and /ɪ/ in unstressed word-internal syllables (this has been called the "weak vowel merger"). Sometimes the merged vowel is transcribed as [ɨ].

A "schwa" before an /n/ is often realized as a "syllabic nasal consonant", which can be transcribed as [n̩].

Here's a previous answer post I wrote that has some more details: When should I use /ə/ or /ɪ/ and why does it seem like they're not used correctly?

Here are some additional comments about IPA transcription, since some comments and other answers have brought up that topic. IPA can be used either to represent certain phonetic vowel qualities (in this use, it is an approximation, since vowels differ along a multi-dimensional continuum but IPA letters are discrete), or to represent certain distinctive features of elements of a language's sound system (this is a "phonological" transcription). In generic transcriptions of the kind you find in a dictionary, it doesn't necessarily represent your pronunciation, or the pronunciation that you are used to hearing; it might represent a conventionalized standard or the way that the word used to be pronounced. So there is potentially a lot of space between an IPA transcription and the actual phonetic form of a word's pronunciation. The use of the symbol "schwa" in a transcription could mean "this word is pronounced with a phonetically mid-central vowel", or it could mean "this word acts in the phonological system of (some variety of) English like it has a fully unstressed vowel in this position", or it could mean "in the most influential historical accents of English, this word was pronounced with a phonetically mid-central vowel".

  • 2
    A little out of scope of question, but isn't the IPA's purpose to be a way to write words the way they are pronounced? – Binyamin May 10 at 12:53
  • 8
    @Binyamin IPA can be used phonetically, to show "actual" pronunciations, or phonemically, to show broad categories of pronunciations. – Mark Beadles May 10 at 13:23
  • 5
    If you look at a diagram of the English vowel phonemes, you see that there are tight contrasts between tense and lax vowels in the front and the back (/i/ ~ /ɪ/, /e/ ~ /ɛ/, /ɔ/ ~ /o/, /ʊ/ ~ /u/), but no contrasts at all in the center -- just /ə/. That means that the entire central vowel region consists of allophones of /ə/, enough to fit into any environment. This dovetails nicely with the necessity to reduce unstressed syllables in a stress-timed language. – jlawler May 10 at 13:36
  • 2
    @Binyamin I'm no expert in this field. But given that this answer talks about American accents, I assume that there also are accents where the schwa sound does sound the same for these words, making the assumption about IPA being consistent about pronunciation true. I can imagine that it's impossible to list every IPA for each accent that exists and they decided to go for a default IPA or something. fwiw, for myself the schwa in about, salad and mountain do sound the same, although I'm not a native English speaker. – Ivo Beckers May 10 at 14:29
  • 4
    @IvoBeckers even then, phonemically the schwa is one phoneme which might be realized as different phones depending on context (syllable location, adjacent consonants, dialects, etc). IPA can notate phonetic distinctions as well when needed, but the point of phonemes is that in general the distinction between allophones doesn't matter – eques May 10 at 15:45
2

I am sure opinions about this will differ from mine, but in my view, a schwa is always a schwa -- a specific sound articulated in a specific way. It is that sound articulated when a vowel is pronounced without raising the tongue from its neutral position, or lowering it, or backing it, or fronting it, or rounding the lips, or modifying it with any other secondary articulation.

I don't doubt that very often phonologists will transcribe a schwa when they hear a vowel somewhat higher or fronter than a real schwa. I myself am not above doing this. But it's a mistake, or at least remarkably sloppy. We want to be studying the actual sounds of a language, not linguists' transcriptional practices. The use of the schwa phonetic symbol as a sort of cover symbol for several different sounds is a fact of linguists, not a fact of language.

1

Greg Lee's observation is on target. The question posed presupposes without argument that the unstressed vowels of 'salad' and 'about' are [ə]. That is classic question-begging, at least for most American dialects. The standard procedure for phonemic analysis is to analyze the distribution of surface segments of a phonetically-transcribed corpus and determine whether certain segments are contextually-predictable and are in complementary distribution with some other segments. We don't know by divine authority that the unstressed vowels in 'about, abut, upper, salad' etc. are phonemically "schwa". We do know that in about, abut the first vowel phonetically matches IPA standard for [ə]. It is entirely credible to phonemicize the unstressed vowel of 'salad' as [ɪ], just as one would phonemicize the unstressed vowel of ignore as [ɪ]. It is also well-known that [ə] and [ʌ] are in complementary distribution (the latter is only unstressed)

There is a phonological rule neutralizing unstressed vowels in certain positions, which leads to vowel quality neutralizations associated with stress changes: for example [ˈtɛlVgræf], [tVˈlɛgrVfi]; [ˈdɛmVkrat], [dVˈmakrVsi]; [ˌkætVˈstrafVk], [kVˈtæstrVfi]. The realization of that vowel V varies in American English between [ɪ ɨ ə]. There is no phonological argument that the output of reduction must be treated as [ə], as opposed to [ɨ], [ɪ], or [ʌ]. IMO the prior question is "what vowel is in this token of this word?", so that we face the individual variability of the initial vowel of [əˈtamɪk] ~ [ɨˈtamɪk] and [tɪˈlɛgrəfi] ~ [tɨˈlɛgrəfi]. But what is the phonemic value of the vowel in these words? It's obvious, in these examples: /æ/ versus /ɛ/. In 'solid' we can tell that the unstressed vowel is /ɪ/ because of 'solidity'. Since there hardly ever or never is actual [ə] is 'salad' (it's usually [ɪ]), there's no reason to posit abstract 'schwa' in that word.

  • So how would you phonemicize the first vowel in about, as /ə/ or /ɪ/? What about comma? (assuming we're talking about accents where salad and valid rhyme) – Nardog May 11 at 23:47
  • @Nardog, I would give them their phonetic values, which in all 3 cases is [ə], and yes, salad and valid rhyme. I don't adhere to the credo of phoneme-saving, so I don't have to analyze schwa out of existence. – user6726 May 12 at 0:12
  • @sumelic, I don't quite understand your point. Are you denying that the first syllable of telegraphy is /ɛ/ (bearing in mind telegraph): if so, we're not talking about the same enterprise. I didn't give the full set of alternations, so yes tense vowels also reduce. – user6726 May 12 at 0:14
  • Yes, I don't really feel like the vowel in the first syllable of telegraphy is /ɛ/. I know that "telegraph" exists and is a related word, but I don't see why that means that they have to have the same vowel phonemes in the first syllable. "photo" and "photography" are also related, but I don't see how they could be said to have the same vowel in the second syllable. "photograph" is related to both "photo" and "photography": would you say that the pronunciation [fotəgræfɪk] (broad transcription) is phonemicized as /fotɑgræfɪk/? – ewawe May 12 at 0:16
  • Yes, the underlying form of the root is /fotɑgræf/, and the suffix is /ɪk/. So as I say, I don't think we're even talking about the same enterprise. – user6726 May 12 at 0:21
1

Even phonetically, the schwa has stood for any "obscure" or "neutral" vowel that is not attributable to a particular peripheral ("cardinal") value for much of the IPA's history.

From Aim and Principles of the International Phonetic Association (1904: 9):

ə stands for any obscure vowel, like the first of English again. When it is found useful to distinguish two such vowels, the closer one may be written ə and the opener ɐ.

From The Principles of the International Phonetic Association (1949: 7):

The Association also recommends that the letter ə be employed to denote any unrounded vowel situated in the interior triangle (see Fig. 4). If a language contains two unrounded central vowels, it is recommended that ə be used to denote the closer one and ɐ for the opener one. Occasionally the form ɜ may be employed to represent another variety of central vowel.

Fig. 4:

It was only in 1993 that [ɘ ɵ ɜ ɞ] came to represent the current values, and they are still little employed as defined in the current IPA.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.