IPA query

Tremmert   Sunday, October 03, 2004, 17:50 GMT
I'm slightly confused by the IPA vowels. In British English dictionaries I've always see the 'e' in 'get' referred to as an epsilon (backwards 3) but when listening to the vowel as recorded on the page:

http://hctv.humnet.ucla.edu/departments/linguistics/VowelsandConsonants/course/chapter1/vowels.html

The epsilon sounds more like a German 'umlaut a', and the vowel which looks like 'e' sounds closer to my pronunciation of the 'e' in 'get'.

I should have a South African accent, having lived here all my life. I thought that the way we pronounce 'e' was the same as RP, but either I'm wrong, or I misheard the recorded vowel, or they aren't recorded accurately. Could you please give me your opinion? My main goal in attempting to learn some of the IPA is to facilitate learning of new pronunciations, so I'd like my knowledge to be correct from the beginning.
Mxsmanic   Sunday, October 03, 2004, 20:49 GMT
The IPA is the same everywhere and for every language; that's the whole idea.

The epsilon character in the IPA represents the vowel sound of 'e' in "bed" in English or "belle" in French.

Some British and American dictionaries and scholars incorrectly use the letter 'e' to represent this sound, but in the IPA, that's a much closer vowel, the one heard at the end of "résumé." These same dictionaries misuse the IPA in other ways as well.

I couldn't listen to the example on the Web page you indicated because I have nothing that plays AIFF files, so I can't comment on that.

The IPA is still the best way to study pronunciation.
Le loup et l'agneau   Sunday, October 03, 2004, 20:58 GMT
"I have nothing that plays AIFF files"

Download Quicktime at:

http://www.apple.com/quicktime/download/
Jim   Monday, October 04, 2004, 00:00 GMT
I wouldn't say that it's incorrect to use [e] for the "e" sound in "bed". Indeed, if we are to trust the Macquarie University's web page then [e] is the best IPA symbol for this vowel in an Australian or New Zealand accent. Note also that in either of these accents the vowel at the end of "résumé" is [æI] not [e] (unless the Aussie or Kiwi is trying on the French pronunciation).

Mxsmanic is right to say that the "e" in "bed" is closer to the vowel represented by the IPA's epsilon (let's call it [E]) if he's talking about a North American accent. I don't doubt that the final vowel in his pronunciation of "résumé" is close to the IPA's [e]. I wouldn't even argue if he said that this is more correct ... after all it is a French word.

But you're not asking about the Aussie/Kiwi pronunciation nor are you asking about the North American pronunciation. You want to know about RP. It seems that the RP's version is somewhere between [e] and [E] (if we trust Mac Uni). So is it incorrect to use one symbol over the other?

Considering the variation in accents around the English speaking world I think it's pretty damn silly to try a phonetic transcription at all (except perhaps for foreign words). I'd much rather see dictionaries use phonemic transcriptions.

Here are the links to Macquarie University's website.

http://www.ling.mq.edu.au/units/ling210-901/phonetics/ausenglish/auseng_vowels.html

http://www.ling.mq.edu.au/units/ling210-901/phonetics/vowelgraphs/AusE_Monophthongs.html

http://www.ling.mq.edu.au/units/ling210-901/phonetics/vowelgraphs/NZE_Monophthongs.html

http://www.ling.mq.edu.au/units/ling210-901/phonetics/vowelgraphs/USE_Monophthongs.html

http://www.ling.mq.edu.au/units/ling210-901/phonetics/vowelgraphs/RPE_Monophthongs.html
Tom   Monday, October 04, 2004, 00:50 GMT
As far as I know, the epsilon is what is used to transcribe American English. In British English, [e] is the best match.

Remember that the IPA does not have infinite granularity.
Sounds produced by humans form a continuum. When you try to describe them with a finite number of symbols, each symbol has to correspond to a RANGE of sounds.

Therefore:
1. The British [e] does not have to sound exactly like the Polish [e].

2. The vowel in British "bed" may be considered a borderline case between [e] and the epsilon. Some linguists will prefer to denote it with [e], others will prefer the epsilon. It's silly to argue about this, as it is a matter of convention.
Jim   Monday, October 04, 2004, 01:05 GMT
Hear! Hear!

Yes, it is silly. I agree. Of course the IPA does have diacritics to further narrow the range of the symbols but, as Tom says, it's still no infinite: there are only a finite number of diacritics too.

http://www2.arts.gla.ac.uk/IPA/fullchart.html
Jim   Monday, October 04, 2004, 01:16 GMT
Typo: "... it's still not infinite ..."

P.S. there are 184 IPA symbols (unless I've miscounted them).
Mxsmanic   Monday, October 04, 2004, 07:11 GMT
The question is which symbol to use for prescriptive pronunciation keys. If you are teaching a non-native speaker how to speak English, is it better to teach him [e] for the sound in "bed," or [E]? I think that [E] is closer to the most common allophones of the phoneme and is safer. Certainly it is just about right on the money for GAE, and it works fine for RP.

There are other serious errors in British and American sources that are even harder to defend. Many dictionaries represent /i/ as /i:/ and /I/ as /i/. This is a very serious error because a difference in the position of the vowel is being transcribed as a difference in length. The phonemic distinction between these two vowels (in "bean" and "bin," respectively) is in position, not length—vowel length is not phonemic in English. An ESL student who believes what he sees and pronounces one vowel as [i:] and the other as [i] will not be understood.

Many English users of the IPA have apparently never studied any other languages and have no clue that their use of the IPA is so seriously in error. But when you are teaching English to speakers of other languages, you need transcriptions that are correct as possible, and not some "English version" of the IPA that will only confuse speakers of other languages who have learned the real thing.
Mxsmanic   Monday, October 04, 2004, 07:13 GMT
Sorry, le loup, but Quicktime makes a mess on my system if I install it, so I guess I just won't be able to hear those sound bites.
Tremmert   Monday, October 04, 2004, 10:19 GMT
Windows Media player seems to work fine.
Tremmert   Monday, October 04, 2004, 10:24 GMT
This website gives [E] for the British pronunciation and [e] for the American. So even if the IPA is standard, it doesn't really help if most people misuse the standard...

http://www.omniglot.com/writing/english.htm
Justin   Monday, October 04, 2004, 14:30 GMT
The IPA symbol [e] should be chosen to represent that vowel in the South African accent.

The RP "e" used to be closer to [e], but in recent years, it has become more open. So the RP "e" now tends to be closer to [E]. In my ears, "e" in General American English is differentiated from the RP "e" due to its more centered quality, not because the American "e" is more open.
Tom   Monday, October 04, 2004, 15:21 GMT

"I think that [E] ... is just about right on the money for GAE, and it works fine for RP."

I will keep using [e] for British English, since this is what is used by British phoneticians. All of my British dictionaries (Oxford, Longman, Collins) use the [e]. So does Daniel Jones's well-known "English Pronouncing Dictionary".
I don't see any reason to go against the convention.

"Many dictionaries represent /i/ as /i:/ and /I/ as /i/."

Oxford, Longman, Collins and Cambridge all use /I/ and /i:/. I suspect you are thinking about very old dictionaries or dictionaries from second-rate publishers.
Tremmert   Monday, October 04, 2004, 20:30 GMT
The Concise Oxford English Dictionary Tenth Ed. uses /bEd/ for 'bed'. It also uses /i/, but apparently mainly for final 'y' in words like 'happy'. Otherwise it seems to use /I/ and /i:/.

My problem was rather the opposite of the common one, since I was attempting to learn the API given my knowledge of English pronunciation. I assumed I'd already know English vowels like short 'e' from dictionaries but since this dictionary refers to a slightly different accent ... it's a good thing I listened to those sound files. Obviously phonetics is one of those things you can't learn entirely from a printed resource.
Jim   Wednesday, October 06, 2004, 03:56 GMT
[E] may well be "just about right on the money for GAE," and may well work "fine for RP" but it's no good for and Australian accent and for a New Zealand one it's probably going to be disastrous (sounding too much like their /@/ as in "cat").

Which symbol to use for prescriptive pronunciation keys? If you ask me, they'd be better off using a phonemic alphabet (like Pitman's ITA or even Shavian) than blundering on with the IPA. Else you have the problem that Tremmert experiences: attempting to learn the IPA your knowledge of English pronunciation which will inevitably be influenced by your own accent.

If you are teaching a non-native speaker how to speak English, never assume that they are familiar with the IPA in the first place. If you're going to use the IPA, use the symbol which best fits your accent otherwise the student will become confused as to what the symbol means. Make sure to explain to them that yours is not the only accent in the English speaking world. However, I think it's best to use a phonemic alphabet.

There are other serious errors in British and American sources that are even harder to defend. Many dictionaries represent /i/ as /i:/ and /I/ as /i/. This is a very serious error because a difference in the position of the vowel is being transcribed as a difference in length.

The phonemic distinction between the two vowels in "bean" and "bin" is both position and length in many dialects including mine. Here's a transcription of my pronunciation: "bean" = [bi:n] and "bin" = [bIn]. The vowel [i] is found at the end of the word "happy" = [hæpi] (in my accent).

"... vowel length is not phonemic in English." Bollocks. It may not be phonemic in Mxsmanic's dialect but it certainly is in mine. Here's a perfect example: consider the minimal pair "some" vs. "psalm"; in my accent the difference is the length of the vowel. Here's how I pronounce "some": [sâm] and "psalm": [sâ:m]. We can't use the IPA here so I'm using the symbols below instead.

________Front________Central__________Back

Close------=y========ï=ü==========í=u
-------------\\--------------||-----------------||
--------------\\-----I-Y------\\---------U-----||
---------------\\-------------||---------------||
Close-mid-----e=ø======ë=ö========é=o
-----------------\\------------\\--------------||
------------------\\------------ê-------------||
-------------------\\------------\\-----------||
Open-mid--------E=œ======3=ß=====^=ô
---------------------\\-----------\\----------||
--------------------æ\\-----------â---------||
-----------------------\\----------||--------||
Open-----------------a=Œ==========å=õ

Where symbols appear in pairs, the one to the right
represents a rounded vowel

I agree'd that "when you are teaching English to speakers of other languages, you need transcriptions that are correct as possible," but they need not be IPA transcriptions. Indeed, phonemic transcritions are better than phonetic ones. So, no, don't use some "English version" of the IPA that will only confuse everyone; ESL students, native speakers, those who have learnt the real IPA and those who haven't alike; use a phonemic alphabet. In a phonemic alphabet /e/ is the most sensible symbol for the vowel in "bed" (unless you're into Shavian, runes or something).