The use of IPA in learning correct pronunciation?

I've had some experiences with "IPA", or to be more exact, some broad transcriptions and custom transcriptions in certain dictionaries.

How should one use the IPA to learn English? I'm sort of leaning toward Cambridge Advanced Learner's Dictionary (the same e-dictionary suggested by admins here), but I still don't understand how Cambridge's symbols correspond to X-SAMPA or IPA (broad), etc.

I personally find it too tedious to check a paper dictionary, unless I can't really find any entry in every electronic dictionary. I'd like to know if many learners here use IPA, and if you try to pronounce a lot of words correctly.

Here's my idea:
If I were to present English to my Chinese friends/relatives, esp. those who don't even know a single English word, shall I be doing something as preparing parallel texts for them? Like in your home country, we also have lots of textbooks for English (and rather old-fashioned).... but audiobooks aren't commonplace yet.

If I do (and I do) parallel texts, I'd hope it is:

- 2 columns, English and your native language
- Use an e-dict, like that of Cambridge, to copy transcriptions. As you move on (using a nice textbook or whatever; I guess someone should have tried Assimil English before?), you can copy less transcriptions in your .doc/.rtf files

I don't care much about the electronic format. Some people say "interlaced" bilingual texts are better, well, that's just me (you may think differently, don't you?)

The main reason of copying the symbols NEAR the text is that I personally forget transcriptions almost as soon as I close the e-dict. I'm doing this for French, and while it's a bit boring, copying has enabled me to understand phonemes much better with my knowledge of IPA (and relevant transcriptions for individual languages, like French).

<<I'd like to know if many learners here use IPA, and if you try to pronounce a lot of words correctly. >>

By pronouncing words correctly, do you mean the standard RP pronunciation? In reality, there are a huge number of almost unintelligible dialects around the world, all of which would have differemt IPA strings.

''If you're trying to learn American English, you should go with an American dictionary.''

American dictionaries don't use IPA and aren't too consistent: for example
they list only /kQt/ under ''caught'', but /kEtS, kQt or kAt, kQt or kAt/ under ''catch (caught caught)''.

The best dictionary for pronunciation are preonunciaiton dictionaries;
both Longman pronunciation dictionary (by JC Wells) and Cambridge Pronunciation dictionary (by Daniel Jones) are good.

You should stay away from the electronic dictionaries with sound filmes, many times the pronunciation indicated/written does not correspond with the voice: for example: horrible: written pronunciation: /'hor@b@l/ but prounounced as /'hAr@b@l/ by a voice.

>>You should stay away from the electronic dictionaries with sound filmes, many times the pronunciation indicated/written does not correspond with the voice: for example: horrible: written pronunciation: /'hor@b@l/ but prounounced as /'hAr@b@l/ by a voice.<<

Part of this is that the pronunciation written down is probably by someone else than the voice, and is not off what the speaker says but rather specific standards or ranges of different dialects. Also, the speaker likely is naturally saying how they would carefully pronounce the word in question rather than trying to "read" the transcription written down. And the matter is that whoever they have speaking likely does not really *exactly* match the forms that the transcriptions are being based off of.

IPA or any other transcription system is useful. Don't rely on transcriptions too much though, since no dictionary gives you accurate transcriptions. For example, no dictionary makes a difference between the two schwas in "Nebraska", as far as I know. And no dictionary lists both "stay-shun" and "stay-shin" for "station". There are too many dialects and ways of speaking, you can't expect to get perfect transcriptions.

Often, when I forget how a word is pronounced, I check in the dictionary to see where's the main stress and what vowels should be reduced. The most reliable one for American English is Merriam Webster, in my opinion. It doesn't use IPA though, but its transcriptions are way better than any other dictionary I've tried (I noticed that most dictionaries that use IPA adapt it to their needs and so use it "wrongly")

I'm afraid I should have said "...I check in the dictionary to see where the main stress is". :(

>>IPA or any other transcription system is useful. Don't rely on transcriptions too much though, since no dictionary gives you accurate transcriptions. For example, no dictionary makes a difference between the two schwas in "Nebraska", as far as I know. And no dictionary lists both "stay-shun" and "stay-shin" for "station". There are too many dialects and ways of speaking, you can't expect to get perfect transcriptions.<<

Another thing is that one cannot expect others to have pronunciations are much at all like those that are given in dictionaries, especially in everyday speech. For me at least, while my *careful* pronunciations are relatively close to GA pronunciations like those one would see in dictionaries to not be problematic, my everyday speech may often be of only limited intelligibility to those solely familiar with pronunciation strictly as according to dictionaries. Probably the largest factor in such, aside from the usual vowel changes, is widespread elision of non-initial /t/, /d/, and /n/ before vowels, when the following vowel is unstressed or when word-final. Such is something that would probably be easy to get used to in Real Life from exposure to such, but would almost certainly limit understandability by people only familiar with dictionary pronunciations.

>>And no dictionary lists both "stay-shun" and "stay-shin" for "station". There are too many dialects and ways of speaking, you can't expect to get perfect transcriptions.

Cambridge lists: /"steI.S@n/
I don't get what you mean.

Well, obviously, I also know there must be some variation in pronunciation of other languages (and my native language, too)...but it doesn't matter that much. I think transcriptions, if any, are enough, as long as they aren't disorganized.

>>I noticed that most dictionaries that use IPA adapt it to their needs and so use it "wrongly"

How? What you mean by "most"?

While this might be 'true', or to put it better, this might be of some importance, after having listened to a lot of bad English of people from everywhere... my own dogma is simple: yes, if a set of transcriptions can be said to be incorrect, it might not be reliable at all, but I'd insist that I LISTEN, preferably from an anthentic source, native materials like audiobooks and whatever broadcasts, to somehow confirm the pronunciation.

What I wanted to say is that transcriptions in dictionaries are not technically perfect. For example, I think most American dictionaries would say "butter" is /b@t@r/, where the t is not showed as a tapped t, and the final sound is not really @ + r, but it's a kind of colored schwa. As for "station", you get /steIS@n/, but no one tells you what kind of schwa that /@/ in there is. It seems to me that reduced sounds before /m/ or /n/ tend to shift toward an /I/ sound.

But I am a non-native speaker, so that's just my opinion. I am not an expert, I don't even know the right terms, but I like learning about pronunciation and I like accents anyway.

So I basically use transcriptions a lot, but only as general guidelines (I mainly check the main stress and the kinds of vowels in a certain word, then I pronounce it according to my personal accent). For example, even if the dictionary tells me to pronounce the t in international, I take it away anyway.

For starters, much of said transcriptions are phonemic, not phonetic, and also conservative. Take the transcriptions /ˈbətər/ for "butter" and /ˈsteɪ̯ʃən/ for "station". Such are actually correct *phonemic* transcriptions for conservative General American. They are in no way contradictory with the actual realizations in such, which would likely be something like [ˈbʌɾ̥ɚː] for "butter" and [ˈsteɪ̯ʃn̩ː] or [ˈsteɪ̯ʃɨ̃ːn] for "station" (assuming vowel length allophony similar to my own, non-merger of intervocalic /t/ and /d/, and [ɨ̃n] for unstressed /ən/).

>>For example, I think most American dictionaries would say "butter" is /b@t@r/, where the t is not showed as a tapped t, and the final sound is not really @ + r, but it's a kind of colored schwa. As for "station", you get /steIS@n/, but no one tells you what kind of schwa that /@/ in there is. It seems to me that reduced sounds before /m/ or /n/ tend to shift toward an /I/ sound.

Exactly. Some might well say a transcription scheme is half-baked, but some might be ignoring the fact that many of them, if not all, are actually "home-baked". If you compare an English and, say, a German dictionary, you might be curious to see they use the same r, but the phoneme r in both languages are obviously different for the absolute learner who has attended a few English and German lessons. Dictionary writers, or writing "groups", often use their own scheme, a set of symbols in their own right, to transcribe A particular language in THEIR particular way. I won't blame them for using some seemingly confusing things such as @ + r for a colored schwa.

I think of a very good example, namely the invention of pinyin. Obviously, it had to be invented to teach natives and foreigners alike, but it's also intended as an auxiliary phonetic alphabet. It's not perfectly phonetic (*some quirks with the spelling), but it's also rather short of exceptions (which natives won't even be aware of anyway). When you read a Chinese-English dictionary, you can of course always see characters along with how their are read; but this is only part of your student world where you see "after all, this weird character is read like this!" with a big grin. In the real world, however, natives are making native mistakes. What you hear is authentic, native, but may not be exactly what the dictionary tells. The same for tones. I don't listen much. I pick up tones like second nature, but without profound exposure, I wouldn't be able to distinguish between the many "shades" of tones as a result of different voice tones, moods, genders, situations, idiomatic usage.... and so on.

>>For starters, much of said transcriptions are phonemic, not phonetic, and also conservative.

Phonemic?

I'd still say it's just a tool, something auxiliary, just like a dictionary, where it's usually found by the average Joe. The learner still has to listen to native speech, exploit the natural talent of picking up new speech to the full and learn.

==

Then, I have another question. Do you find transcriptions - in any organized form - to be useful? If a learner can bear using just bilingual texts with audios of native speech to learn everything from scratch, like me, then it might be good to put transcriptions -- the target language text, and below it, the transcriptions.

for English, I may use my own Cambridge transcriptions (or even tweak it a little for my own understanding); for some others, I do the same depending on how they "sound" like. Interestingly, while Assimil only uses a rather rough and inconsistent pronunciation guide (short of being transcriptions), its guide does serve, for me, like transcriptions.

>>Exactly. Some might well say a transcription scheme is half-baked, but some might be ignoring the fact that many of them, if not all, are actually "home-baked". If you compare an English and, say, a German dictionary, you might be curious to see they use the same r, but the phoneme r in both languages are obviously different for the absolute learner who has attended a few English and German lessons.<<

The matter is that for transcriptions of languages at the phonemic level, it is commonplace to just use /r/ for the rhotic consonant in those languages if those languages only have one rhotic consonant phoneme (as is the case with both English and German). Such is not ambiguous, as it seeks to represent phonemes and not sounds to begin with, and because there is nothing to confuse such with in such languages (unlike in, say, Spanish, which has two rhotic consonants).

>>Dictionary writers, or writing "groups", often use their own scheme, a set of symbols in their own right, to transcribe A particular language in THEIR particular way. I won't blame them for using some seemingly confusing things such as @ + r for a colored schwa.<<

Actually, representing the sound [ɚ] as /ər/ phonemically is not inaccurate at all. Rather, it is quite clear that there is no independent phoneme /ɚ/ which contrasts with /ər/ in rhotic English dialects, and that /ər/ is an accurate representation of the underlying form that results in [ɚ]. This is a whole matter in itself, though, and I am not going to get further into this unless you really are interested in it specifically.

>>For starters, much of said transcriptions are phonemic, not phonetic, and also conservative.

Phonemic?<<

You should go read this:

http://en.wikipedia.org/wiki/Phoneme

You really have to wrap your brain around this before much of what I am saying here in this discussion will really make sense, honestly. The distinction between phonemes and phones (sounds) is a fundamental matter in modern linguistics which one really must understand to understand how languages work in reality.

>>Then, I have another question. Do you find transcriptions - in any organized form - to be useful? If a learner can bear using just bilingual texts with audios of native speech to learn everything from scratch, like me, then it might be good to put transcriptions -- the target language text, and below it, the transcriptions.

for English, I may use my own Cambridge transcriptions (or even tweak it a little for my own understanding); for some others, I do the same depending on how they "sound" like. Interestingly, while Assimil only uses a rather rough and inconsistent pronunciation guide (short of being transcriptions), its guide does serve, for me, like transcriptions.<<

When I am reading about some given language, phonetic transcriptions will give me far more information about how it is really pronounced than any subjective descriptions of such, even though such may be misleading in and of themselves (as because people do not always speak the same way all the time, and such often do not take much dialect variation into account). Likewise, phonemic transcriptions are useful in helping separate the actual underlying forms in a given language from the realized surface forms, and help give an abstract view of a language separate from the particulars of the actual phonetics of it. Unfortunately, the two are very often confused, and whether something is a phonetic or a phonemic transcription is often not sufficiently specified, which can be problematic.

>>Such is not ambiguous, as it seeks to represent phonemes and not sounds to begin with, and because there is nothing to confuse such with in such languages (unlike in, say, Spanish, which has two rhotic consonants).

So, how is Spanish represented in a dictionary?

>>Rather, it is quite clear that there is no independent phoneme /ɚ/ which contrasts with /ər/ in rhotic English dialects

So, this is a phonemic scheme? That's interesting. The moral of the story would then be a serious student would find it very helpful to study a bit of phonetics - at least to enable himself to identify all phonemes, work on pronunciation, and get it right...

==

There has been discussions elsewhere (in another famous, 'comprehensive' forum), and some (mainly English speakers) tend to hold a view that they think it might be (since they would never be sure) beneficial to use IPA for learning English, but it might not for things like Mandarin. Given the special things about Mandarin, as a half-native, I actually do believe IPA helps. The foreign learner won't normally know any Chinese beforehand, and so they can't really learn Mandarin in "Chinese".

Hm... I might have to apologize that I should have posted this in the Languages board. I should have asked whether IPA is _useful_ for _other_ languages as well (and how useful you find it).

>>The distinction between phonemes and phones (sounds) is a fundamental matter in modern linguistics which one really must understand to understand how languages work in reality.

I know a smattering of concepts about this. I was just wondering how "phonemic" a particular transcription scheme could be said to be. I've got your point. :)