Phonemic transcription

What kind of transcriptin should be used in the teaching of English at secondary schools?

English as a first or a second language?

More often what you'd tend to find in texts for ESL students is the International Phonetic Alphabet. Most dictionaries these days use the IPA ... or maybe you could say that they abuse it. Dictionaries and textbooks which use the IPA cannot help but give preference to a particular accent. This can give learners the false impression that there exist such a thing as a standard English dialect.

One problem with a phonetic alphabet is that it tends to give too much detail. The same word may be transcribed with different phonetic letters depending on the accent your transcribing. This might not be such a problem for native speakers but for those who are learning English as a second language this detail may be confusing.

To try avoiding this you could use a phonemic alphabet like the Initial Teaching Alphabet. Most secondary school students who are native speakers of English would be beyond this alphabet. The ITA could be used at primary level for native speakers. The problem with the ITA is that it isn't that popular so your students would learn it from you then would have little use for it in later life.

Sites about the IPA:
http://www.arts.gla.ac.uk/IPA/ipa.html
http://www.antimoon.com/how/pronunc-soundsipa.htm
http://www.omniglot.com/writing/ipa.htm
http://www.wikipedia.org/wiki/International_Phonetic_Alphabet

More info on the ITA:
http://www.itafoundation.org/
http://www.unifon.org/ita-eval.html
http://www.austehc.unimelb.edu.au/asaw/biogs/A000907b.htm

An interesting article:
http://www.wikipedia.org/wiki/Phonics

i think ita seems a bit cheesy...

Yeah, I can see that but are there any phonemic alphabets for English which aren't as cheesy?

How about the Antimoon ASCII alphabet?
http://www.antimoon.com/how/pronunc-ascii.htm

to tom, about http://www.antimoon.com/how/pronunc-ascii.htm.

e: -> 3: (1) turn, learn
Ou -> ..u (2) go, home
TH -> dh (3) this, mother

i think your ascii alphabet looks interesting, though it lacks consistency. the most important mistakes i found, would be those:

1 - each sound should have a unique symbol, and you are already chose "e" on the representation of another sound. using "3:" (three:) instead could solve this problem.

2 - if you already chose ".." for the "schwa" sound, you should be consistent and use it when always applicable. if you don't like the idea of writing /g..u/ (go), you should have thought about it when you first made your choice. also, if you thought of it because of the american variation of that sound, then you should have put the equivalent ipa symbol on your table, avoiding confusion.

3 - like tS and tZ, there should be th and dh, since the second is just the voiced version of th. it's all about consistency.

apart from that, other changes would only be necessary depending on whether or not you want to consider the diffence of some certain sounds, like your "e" sound in bed and in say. or the "i" in hit and heat.

there are clear differences between those sounds, but even a lot of dictionaries have chosen not to mention them, because they think it makes learning the pronunciation easier, and it certainly does.

however, that should be seriously taken into account if you believe your accent is also very important to you, apart from only being able to make yourself understood, since those "leetle" things can make a big difference to your accent.

therefore, changes such as /bEd/ for bed (instead of /bed/), should maybe be taken into account.

"difference" not diffence. uh, typos, whatever...

1. [e] is a different sound from [e:].
2. [Ou] is a symbol that covers both the British and American way. In addition, I don't perceive the vowel in "go" as composed of schwa and [u]. I perceive it as a separate vowel.
3. We thought about [dh], but we decided it would be unintuitive. I can read[THen] much more easily than [dhen].

"bed" has an [e] sound, "say" has [ei].
"hit" has an [i] sound, "heat" has [i:].

Tom,

I've been using your system for a while ... only when posting on Antimoon though ... or at least I though I had been.

Just now I looked again at it and noticed that you use [TH] and [th] for "then" and "think". I don't know if you've changed it since I first looked at it or if I just wasn't looking properly in the first place. Instead of [TH] and [th] I'd been using [D] and [T] like they do in SAMPA and Xenglik*. What does this say about which alternative is more intuitive?

With the [TH] verses [th] convention you have to remember "Capitals for the voiced consonant and lower case for the unvoiced one." This is not intuitive in the least. This is a completely useless bit of information to remember. This convention would have people more often having to refer back to the chart.

The alternative which would have [dh] verses [th] would be better. It's easier to remember which is the voiced one and which is unvoiced because [d] is voiced and [t] is not. However you're still using digraphs which are unnecessary and could lead to confusion. How would you describe the pronunciation of words like "adhere"? I'd go for [D] and [T]. This is avoids the confusion. It's clear, concise, easy to remember and fits conventions used elsewhere.

Mee mentions your [e:]. I too have somewhat a problem with your use of [e:]. It relates to the problem that Mee metioned but it's got a bit more to it. The sounds [e] and [e:] are different, fair enough, but using this notation you give the impression that the difference is only in the length of the vowel. They differ in more than just length. The sound you use [e:] for is closer to a long [..] than a long [e].

The confusion that this could cause aside, my main arguement against your use of [e:] is that it makes it impossible to describe the sound of the vowel you could make by extending [e] into a long vowel. This is no problem, you might be inclined to argue, because the Antimoon ASCII alphabet is meant to be phonemic and there is no such phoneme in English. But is there really no such phoneme?

When I think about how I say the words "bred" and "bread" I sort of tend to think that the [e] is longer in the second word. When I say the word "yeah" it seems there is this same long vowel here too. According to the department of linguistics at Macquarie University, Sydney, Australia* "there is evidence that some centring diphthongs may be ... in the case of Australian English, changing to monophthongs". It seems that Aussies are tending replace the [e..] with a long [e] in words like "where" and "air". I'd side with Mee in his support of [3:] for the vowel in "turn", "firm", etc. leaving [e:] free for the long [e].

It seems you've made steps to cover both the British Received Pronunciation (RP) and the General American (GA) accents but haven't you left everyone else out in the cold? The "o" sound in "go" comes out close to [..u] in RP whereas in GA it's more like [ou] so you use [Ou] to cover both.

This courtesy is not extended to other dialects. The "ai" sound in "bait" comes out closer to [@i] in Australian and New Zealand accents however you stick with [ei] to represent this diphthong instead of something more fair, like [Ai] for example. [fee g^u m@it] Similarly, the "ou" sound in "out" could come out as [@o] rather than [au] in the Aussie accent.

I don't understand the need for the "u" in [Ou]. Sure, it tells us that we're dealing with a diphthong but if it were up to me I'd just go for [O] by itself to represent the whole diphthong. [O] by itself could do the job and using it would reduce the clutter. Similarly I'd go for [A] to represent the "ai" sound in "bait".

There are also phonemes missing. There's no gutteral stop, as in "ah!" and in the Cockney "butter". You could use [H] or [?], ie [^H] or [^?] & [b^H..] or [b^?..]. There is no sysmbol for the "ch" in "loch". You could use [K], i.e. [loK]. Some people pronounce the "wh" in "which", "whale", etc. as a different phoneme, not just [hw] but like an unvoiced [w], you have no symbol for this. You could use [W], i.e. [WitS] & [WAl]

* See these sites:
SAMPA
http://www.phon.ucl.ac.uk/home/sampa/english.htm
Xenglik
http://www.unifon.org/dual-reader.html
The department of linguistics at Macquarie University
http://www.ling.mq.edu.au/units/ling210-901/phonetics/ausenglish/auseng_vowels.html

thanks, jim.

i do think T and D are the best for that. i actually thought about mentioning that as well, and also some other stuff, but i thought if i was too strict, tom would get annoyed, and would just ignore it. well, he wasn't much receptive, anyway.

i agree with you in your whole ideas, jim. though, i think it would be better to add more different symbols, instead of using just one symbol to represent different variations of a sound. that way, each person would be able to use them in the way they consider more apropriate, therefore being able to make justice to their own accent, or even just to use their prefered accent when phonetically representing their speech in the forums.

i don't think making the phonetic represantation simpler is the best way to study languages. many have tried, but everybody agrees that the ipa is the best way to represent the sounds. obviously we're only looking for something else because it's too much trouble to use those in a computer. i'd just like to remind everybody, that what makes the ipa better is it's consistency, but that includes its complications as well. i just think that, to make the best representation of the sounds, we should follow that example.

there's a big difference between chosing how to represent the sounds, and chosing how to use them in the phonetic transcriptions of speech. i think that should be made clear.

and just in case someone's forgot: the ":" symbol, is only used to represent the length variation of a sound, and not to represent different sounds.

i'm not sure whether this is your case tom, but i know that some people from some central european countries, tend to pronounce for instance, the /i:/ in "sleep", the same way as in "slip", while latin language speakers like frenchies, tend to go the opposite way. so many people tend to think it's the same sound, but it simply isn't. obs: jim, i do know that a lot of australians pronounce them with the same sound, only with a different length.

tom, if you don't perceive the use of the "schwa" sound in that case (go), you shouldn't be using the equivalent ipa symbol on your table, then. have you still not noticed? i really don't want to offend you, tom, but to begin with, i think you should undertake a deeper study of phonetics before rejecting other people's suggestions.

"someone's forgotten"

"Instead of [TH] and [th] I'd been using [D] and [T] like they do in SAMPA and Xenglik. What does this say about which alternative is more intuitive?"

-- It certainly says a lot about which is more intuitive for the 0.0001% of the population who know SAMPA and Xenglik.

[wiT], [Den], etc. are completely out of touch with both IPA and the conventional spelling "th". In my experience, students have little difficulty telling [th] and [TH] apart. In any case, would it be any easier to tell [t] and [T] apart?


"The sounds [e] and [e:] are different, fair enough, but using this notation you give the impression that the difference is only in the length of the vowel."

-- True for people who know more of IPA than is typically presented in English dictionaries. In virtually all dictionaries, [i:] is presented as a separate vowel, and not merely a longer version of [i].


"When I think about how I say the words "bred" and "bread" I sort of tend to think that the [e] is longer in the second word."

-- And that is supposed to be the justification for "freeing" [e:]? The argument about the "long [e]" in Australian English is uninteresting, because 1) the ASCII alphabet was not meant to cover Australian English, 2) the long [e] is not used in any English dictionary. (and rightly so)


"I don't understand the need for the "u" in [Ou]."

-- we used [Ou] to make it look like the symbol used in English dictionaries


The glottal stop, [hw], etc. were omitted because they're unimportant.

"tom, if you don't perceive the use of the "schwa" sound in that case (go), you shouldn't be using the equivalent ipa symbol on your table, then. have you still not noticed?"

-- Unfortunately, I have to, because that's the symbol dictionaries use.

Tom,

I couldn't claim to know either SAMPA or Xenglik in any significant detail. I just found them on the web and thought their way of representing [th] and [TH] made a lot of sense. Perhaps my having stumbled upon websites about phonetics and spelling reform has influenced me or perhaps I'd just been following a pattern you (and others) use elsewhere, i.e. [S] for "sh" (as in "ship") with [Z] for the voiced consonant.

Consistant application of the same pattern would, I think, make it easier to tell [t], [d], [T] and [D] apart than [t], [d], [th] and [TH]. If you have [s], [z], [S] and [Z] it follows that you'd have [t], [d], [T] and [D]. If you use [th] and [TH], why not use [sh] and perhaps [SH]? Maybe this would only make sense for people with at least a little interest in phonetics but if you look around the forum you'll notice that very few seem to have enough interest even to use your ASCII alphabet in the first place.

Well, I've got to admit that [th] & [TH] fit better with conventional spelling but I don't think they are any better in touch with IPA. The IPA doesn't use digraphs for [th] and [TH] thus, I believe, [T] and [D] fit better. The arguement against the use of digraphs, that they could lead to confusion, was not best exemplified with the word "adhere" because your system doesn't use [dh]. You could write [adhe..] (RP) or [adher] (GA) without any ambiguity.

However, when it comes to words like "lighthouse" and "hothouse" the problem does arise. How would you transcribe these words? [laithaus] and [hothaus] appear to have [th]s rather than [t]s followed by [h]s. Of course, the problem might not arise that frequently and there are ways around it when it does but if I'd been the one designing the alphabet I'd have avoided the problem all together with the use of [T] and [D].

Your case against most of my other objections seems to be that these considerations are unimportant. To some extent you have a point. Though, there are many topics being discussed on your forum which can get into a great deal of detail. At times topics about dialect crop up and we want a consice way of writing about them. If there are holes in the phonetic system then we have to look elsewhere or bend it to suit our needs.

In a forum about the English language how can you be sure that the need to discuss the various ways English is spoken will not arise? I have found myself wanting to write about these sounds from time to time. I don't agree that these missing phonemes are unimportant. If someone were to ask about "the British accent" how better to set them straight on the fact that there is no one "British accent" than to give them a few examples of the different accents heard in Britian. To do this you might want a symbol for the glottal stop, the unvoiced [w] as opposed to the voiced one or [hw], etc.

I'm dissapointed to think that dialects other to RP and GA are not regarded as worth your consideration. There is more to English pronunciation than RP verses GA. I would want to avoid perpetuating the myth that there exist but two competing standards of English. There are many dialects and it would benifit students to recognise this. I don't agree that the makers of dictionaries are right to omit this either.

I never meant this as an attack against your system. I don't think it's an entirely bad one. One thing that I really like about it is the [.] verses [..] distinction. I just wanted to point out what I see as its short comings. I will continue to use your system nonetheless in my Antimoon posts. When the need arises I'll bend it to my needs or use something else. However, I'm not going to use it, as it is, elsewhere.