Talk:International Phonetic Alphabet for English

Contents

1 Obsolete discussion of title 2 Pronunciation of 'cot' 3 IPA for English? 4 /EI/? 5 IPA template 6 Lot and father 7 Keywords, SQUARE vowel 8 Quick reference chart 9 A minor error. 10 Can we drop SAMPA 11 Affricates 12 What about... 13 Syllabic n or l 14 ʊə sounds 15 Pronunciation guides in Wikipedia articles

Obsolete discussion of title

If it's American English, you might want the title to say that

Well, the intention is to also include British English, I just haven't gotten to it yet :-) Nohat

Will that be Southern British English, Midland British English, Northern British English, Scottish British English, Northern Irish British English, Welsh British English, South Western British English or Geordie ? I ask because they differ from each other radically. If "British" English exists at all, it's a written language not a spoken one. -- Derek Ross

It's Received Pronunciation, the system used in British dictionaries, even if no one actually talks that way. Having the particular IPA symbols for the various dialects of British English, and American English, for that matter, would certainly be an interesting page, perhaps elsewhere on Wikipedia. However, for the most part, phonetically speaking, a direct mapping can be made between the sounds of a dialect and either standard American English (SAE) or Received Pronunciation British English (RP). Additionally, descriptive work on dialects is generally made in terms of a standard dialect. There aren't official standards for English, but there are generally agreed upon standards for SAE and RP, and what I put here I think generally reflects that. Nohat

Hey, my speech is pretty close to RP! ;) What does everybody think of doing SAMPA here as well? It could be confusing have two columns of symbols, but then again, for many users, this page is just lots of "?" ... -- Tarquin 21:01, 5 Sep 2003 (UTC) hmmm... it's here: SAMPA chart for English. even if we don't merge, we should standardize both pages to the same format (and title!) I think. PS - side by side tables are bad!

PPS - I made some OGG recordings of RP phonemes some time ago that we could link here. I think they're linked on the talk page for the RP article. -- Tarquin 08:56, 6 Sep 2003 (UTC)

Pronunciation of 'cot'

Isn't cot (a bed for a baby) pronounced as /kɒt/ in British English? /ɑː/ is the sound as in father, /'fɑːðə/. —Sinuhe 11:09, 13 Aug 2004 (UTC)

Yes, that's correct. -- Tarquin 16:20, 14 Aug 2004 (UTC)

Provided that by "British English" you mean "Received Pronunciation". -- Derek Ross | Talk 18:28, 2004 Nov 1 (UTC)

IPA for English?

Can't the information on this page be located at, say, the phonology of English instead? After all, it would be strange if we had articles for International Phonetic Alphabet for French, International Phonetic Alphabet for Japanese, International Phonetic Alphabet for Mandarin Chinese, etc; instead, IPA is used for those languages to explain their sounds, either in their phonology sections or in separate phonology articles. Moreover, we will eventually have a phonology of English article, which will completely duplicate the information on this page anyways. So why do we need this article? -- [[User:Ran|ran (talk)]] 16:47, Nov 1, 2004 (UTC)

That makes sense. -- Derek Ross | Talk 18:29, 2004 Nov 1 (UTC)

Certainly an article on English phonology would be interesting, but it would also be of a much larger scope than this article. Besides individual sounds, it would discuss syllable structure, intonation, phrase boundaries, cross-word phonetic effects, coarticulation, and all the other topics covered by phonology. This article's value is in the fact that it is simply a list of the IPA symbols used for transcribing English, nothing more. That in and of itself is useful, but it is also useful for readers who don't know IPA and encounter IPA transcriptions of English. This page is linked to from some of those pages already, and it is in Template:IPA. Nohat 21:21, 1 Nov 2004 (UTC)

It's useful to have this page separate because this is the English-language Wikipedia. I think that justifies having this page exist, without necessarily starting pages for French, Japanese, Mandarin Chinese, etc. If and when the phonology of English article is written, it need not make this page redundant. This page deals only with the English pronunciation, not English phonology. --Angr 23:15, 30 Jan 2005 (UTC)

/EI/?

I am a North American speaker, and I think I pronounce the diphthong in play with something closer to [EI], rather than [eI]. Should we rename the phoneme to that, or am I just weird? --Sonjaaa 07:27, Jan 13, 2005 (UTC)

You're just weird. Kidding aside, [eɪ] is a pretty standard transcription used for English for the vowel in play. Given that, the salient acoustic attributes of diphthongs are not the exact starting and stop points of the diphthong, but rather the direction and distance. Therefore, it is likely that in the production of a dipthong like [eɪ] the starting point could vary considerably, and [e] and [ɛ] aren't really very far away from each other in the vowel space. There are, however, solid historical and phonological reasons to stick with [eɪ]. Nohat 01:19, 17 Jan 2005 (UTC)

IPA template

I've added the IPA template to most of the IPA transcriptions in this article - in particular all of those that include special characters (coded with &#) which should make the article readable on many more browsers (including the much derided MSIE). rossb 21:56, 16 Jan 2005 (UTC)

Lot and father

Although there is variation, and the observation does not hold for many from the East coast, the vowels of 'lot' and 'father' are identical for the majority of North Americans: 'father' and 'bother' are a perfect rhyme, and 'bomb' and 'balm' are homophones (provided the latter is pronounced without the "l"). The question is whether that vowel is long or short. British pronunciation dictionaries that include American pronunciation, e.g. Wells' Longman, transcribe it as long. But I think that's wrong for both phonetic and phonological reasons. On the phonetic side, the vowel of "hot" is not significantly longer than the vowel of "hat" in American English, so it doesn't make sense to transcribe the former [hA:t] but the latter [h{t]. (Both are longer than the vowel of "hit", but that's just because low vowels are inherently longer than high vowels.) On the phonological side, long vowels and diphthongs are prohibited before certain clusters in English, including [sp] and [sk]. There are no words in English like [ki:sp] or [tu:sk]. If "wasp" and "mosque" are transcribed [wAsp] and [mAsk], they obey this generalization, but if they are transcribed [wA:sp] and [mA:sk] they seem to be peculiar exceptions. For these reasons I think it's better to treat AmE [A], in both "father" and "lot" words, as a short vowel. --Angr 13:09, 6 Feb 2005 (UTC)

I consider the "a" in "father" as long for several reasons: 1) most dictionaries (including the one you refer to) transcribes it as long, 2) the stress is on the first syllable, making it long, 3) it is apparent to me from pronouncing the word that it is long, 4) I do not notice a difference in length between BrE and AmE. You say that long vowels and dipthongs are prohibited to appear before certain [consonant?] clusters in English, but "father" probably doesn't fit that case.--Tokek 17:34, 7 Feb 2005 (UTC)

Tokek, are you British or American? I don't deny the British pronunciation of "father/start" has a long vowel; that's clear. But to the best of my knowledge, the "father/lot" vowel is transcribed long for AmE only in British-published dictionaries that include American pronunciations, not in dictionaries published in the US. The only American-published pronunciation dictionary I know of that even uses IPA symbols is Kenyon & Knott, and they don't transcribe any vowels as long for General American (though they do for East-Coast accents). Other American dictionaries like Merriam-Webster, Random House, and American Heritage use in-house phonetic symbols (rather than IPA) that are ambiguous regarding length. So it would seem that while the British feel the American pronunciation of "father/lot" has a long vowel, Americans themselves aren't so sure. As for stress, yes, it can make vowels phonetically longer, but that doesn't make them phonologically long. The [I] of "ribbon" is longer than the [I] of "tonic", but that doesn't mean we should transcribe the former as ["rI:b@n], does it? Furthermore, I do think the vowel of RP [mA:sk] "mask" is much longer than the vowel of AmE [mAsk] "mosque", so we shouldn't transcribe them both [mA:sk]. (I admit that's only my impression; I don't have phonetic evidence to back it up.) --Angr 19:21, 7 Feb 2005 (UTC)

Usonian dictionaries don't transcribe as long because the length isn't phonemic. But I speak something pretty close to standard US English, and it's definitely long phonetically. Actually, this vowel tends to be longer than others in most languages that have it, probably because it physically takes longer to drop the tongue that far. But compare father with feather: in feather, the th is ambisyllabic. (That is, it belongs to both syllables, ending one and beginning the other, so it's impossible to decide where to put the syllable break. This is one of the places where English is just plain weird.) You get ambisyllabic consonants after short vowels in English. However, in father the break is clear, fa.ther, implying that the vowel is long: . Of course, it's another question whether you want to bother with that level of phonetic detail. kwami 22:41, 2005 May 16 (UTC)

Yes, low vowels are always longer than high vowels because of the former's greater sonority. But I don't think AmE is significantly longer than the other low vowel, . I'm sure if you recorded several General American speakers saying Say 'hot' for me and Say 'hat' for me and averaged the durations of hot and hat you'd find no significant difference in duration between them. (On the other hand if you compared several Americans' Say 'hot' for me with several RP-speakers' Say 'heart' for me, you would probably find the latter to be significantly longer, which is why it annoys me so much that British dictionaries attempting to show American pronunciations always transcribe American hot as being the same as RP heart.) As for ambisyllabicity, it's impossible to define phonetically; it can never be more than impressionistic. Psycholinguistic experiments designed to determine whether intervocalic consonants are ambisyllabic in English have given contradictory results. Daniel Kahn claimed that flapping applies only to ambisyllabic /t/, but since flapping applies after long vowels (heater ) and r-colored vowels (party ) as well, he must have made different assumptions about ambisyllabicity than you do. (For my part, I don't feel the syllable break in father is any clearer than it is in feather.) --Angr/_comhrá 22:58, 16 May 2005 (UTC)

Keywords, SQUARE vowel

I want to give my reasons for some changes I made that Nohat just changed back. First, I used the keywords FLEECE, KIT, FACE, etc., that are used in John Wells' Accents of English precisely because they are not minimal pairs and therefore there will be no possibility of confusion. The way an Australian says "bad" and the way a person from Northern Ireland says "bid" may both sound to an American like "bed". But "trap" and "kit" won't get confused because there are no words "trep" or "ket". Likewise "dress" is unambiguous because there are no words "driss" or "drass". Second, if we are using the convention of transcribing the DRESS vowel as IPA [ɛ]/SAMPA [E] (rather than [e], which is usual in British publications), then the SQUARE vowel of AmE should be transcribed [ɛɹ]/[Er\] rather than [eɹ]/[er\]. I think the majority of Americans would identify the portion of the SQUARE vowel before rhoticization begins as the same as the DRESS vowel. Of course this transcription predicts that "Mary" and "merry" are homophones, but I think for the majority of Americans, they are homophones. --Angr 06:20, 7 Feb 2005 (UTC)

I've always seen complete listings of English phonemes using near-minimal sets. This makes it clear that the vowels are in fact different. Since we're not listening to actual productions of these words, but reading spellings of them, I'm not quite sure how choosing non-minimal pairs eliminates the possibility of confusion. Maybe a pair of words for each vowel, one from the near-minimal set, and one from your set?

As for the vowel in square, apparently my intuition about those vowels is contrary to the received wisdom about American vowels before R. It's definitely [e] and not [E] for my internal representation of these vowels, but this seems to be a minority grouping. I've never seen this interpretation about these vowels described in any dialectological texts, so I suppose we needn't give it any credence here, but I don't think it's pathological, because my speech is interpreted as "normal" or "Californian" by linguists who are familiar with American dialectology, but who don't know my linguistic background. It's just that the idea that the vowel of square is [E] seems very foreign and "wrong" to me. Maybe it's by analogy with similarly-spelled forms: mate = /met/, maze = /mez/, mane = /men/, male = /mel/, mare = /mer/. On the other hand, I also identify the vowel in words like bank and Jenkins as being that vowel as well, so my intuitions aren't always based on a spelling analogy: sake = /sek/, sac = /s{k/, but sank = /seNk/ not /s{Nk/ for me. Feel free to change it back to /E/. Nohat 21:08, 7 Feb 2005 (UTC)

I guess you're right that as long we're sticking to writing, there won't be confusion. But I still like Wells' keywords. For someone who's never heard of the cot/caught merger it can be confusing when someone asks "Do you pronounce cot and caught the same?" especially if the questioner has merged them. "Do lot and thought rhyme for you?" is much easier to understand, no matter what the dialect of the questioner. Anyway, I'm changing square (now identified as Mary/bared) back to /Er/. And if you merge the vowels of bank and Jenkins, you have some nerve calling that girl "just weird" for saying [EI] instead of [eI]! ;-) --Angr 22:38, 7 Feb 2005 (UTC)

Quick reference chart

In response to a request from a nonlinguist Wikipedian, I've added a quick reference chart in alphabetical order (with special symbols coming after the letters they most closely resemble) where those who encounter IPA transcriptions can look to see what the symbols mean, without knowing what "stop", "fricative", etc. mean. --Angr 21:22, 16 Feb 2005 (UTC)

A minor error.

"* The vowel is pronounced as [ɛɹ]."

Actually, word "bear" is not pronounced with an [ɛ] followed by an [ɹ] but with a raised [ɛ] ([ɛ, in between [ɛ] and [e]). To prove this, speakers from New York city and Philadelphia (and other northern dialects) pronounce marry, merry, and Mary all differently: [mæɹi], [mɛɹi], and [mɛ ɹi]. Thus there must be two ways to pronounce "merry": [mɛɹi] (Northern) and [mɛɹi] (general).

The question is, is General "Mary/merry" pronounced like Phila./NYC "Mary" or like Phila./NYC "merry"? I think probably the two pronunciations of "merry" are the same, and it's Phila./NYC "Mary" that doesn't occur in General. Also, I believe that the American speakers for whom "merry" and "Mary" are distinct are the same as those for whom "can (n.)" and "can (vb.)" are distinct, and that most of them have the intuition that the vowel of "Mary" is the same as the vowel of "can (vb.)", i.e. the vowel I transcribed as in the article. Now that symbol is somewhat misleading since for many people (but not all!) with a contrastive , it's actually a closer vowel than and tends to be diphthongal. Whatever its realization is, let's take as the symbol for the phoneme, which occurs (as a phoneme) only in and around Philadelphia and New York. Then, for people with a three-way "marry/merry/Mary" contrast, these can be phonemicized and "can (vb.)/can. (n.)" can be phonemicized . Then, for speakers of General American (in which at least "merry" and "Mary" have merged, and for many "marry" has too), the phonemic representations are ( for those who have it distinct) and . Does anyone know of any dialects that have the "can/can" contrast but not the "merry/Mary" contrast? (The converse case, with the "merry/Mary" contrast but not the "can/can" contrast, is to the best of my knowledge the situation everywhere outside of North America.) --Angr 20:34, 19 Feb 2005 (UTC)

Aha. You're on the right track. At least you see where I'm coming from, and you are not on of those who simply brushes off any suggestions, stating that the IPA respresentation of English is totally correct. In fact, it is far from correct ("bought" and "bore" using the same vowel sound [O:]? Give me a break).

So, to answer your questions, speakers of modern American English (spoken by somehwere around 65% of adults and 80% of children [just estimates]) use the Philadelphia/New York City pronunciation of "Mary" ([mɛ˔ɹi]) for "marry," "merry," and "Mary." You are absolutely right about the fact that speakers that distinguish those three words also distinguish "tin can" and "someone can" (quite drastically, it amazes me that the dictionary lists "lamp" and "lap" as having the same vowel sound!) The sound in P/NYC uses for "can" (noun) and that MAE uses for both "can" (verb) and "can" (noun) is called "tense A," as opposed to lax A found in "pack." This sound, unfortunately is incorrectly represented in this article as /æ˔/. This insinuates that the sound is somewhere in between [ɛ] and [æ], which it is not. It is actually in between [ɛ] and [e]. Try it yourself with the words "lad" and "land", and "led." Additionally, this tense A sound in "land" is actually a diphthong, as you said, and it must be represented as [ɛ˔ə]. That being said, tense A is the same sound as the A in Mary, just without the final schwa. (By the way, the difference between lax A and tense A is a lot greater than you think; almost everyone notices my "odd" pronunciation of "Spanish" [I use lax A [æ], thanks to my Phildelphia heritage; MAE uses tense A [ɛ˔ə]] which, when I say it, certainly does not rhyme with "mannish.")

So in summary, "marry," "merry," and "Mary" are phonemicized in Philadelphia/New York City as /mæɹi/, /mɛɹi/, and /mɛ˔ɹi/, and in MAE they are transcribed as /mɛ˔ɹi/, /mɛ˔ɹi/, and /mɛ˔ɹi/. P/NYC "can" and "can" are /kæn/ and /kɛ˔ən/, while in MAE they are /kɛ˔ən/ and /kɛ˔ən/. Much of my information comes from a computer program I have written, in which the user types in a string of IPA symbols and word is pronounced. I have verified all my findings with this program. So if you have any other questions about American English, I'll see what I can do. And we have a lot of work cut out for us to settle the "aw" issue; "caught" is most certainly not prounounced [kO:t]. In fact, that's how it is pronounced in British English. Are you saying we pronounce the word the same as the British do? That's absurd. Such a distinction is on of many features people point at in a British accent ("wotuh" for "water," "dotuh" for "daughter").

--Jordan

First of all, please sign your posts! Even if you don't yet have a user name, at least still sign with your IPA address so we know who's saying what! (And if you're planning on sticking around at Wikipedia, please sign up for a user name!)

Secondly, I think you're generalizing your own pronunciation to other people. I had a student from Long Island who once produced for me the triplet "can (vb.) - can (n.) - Ken" and for her, the vowel of "can (n.)" was between that of "can (vb.)" and that of "Ken", i.e. it was more open than "Ken" and more close than "can (vb.)". The only correct transcriptions for her "can (n.)" would be ; since the vowel in question is etymologically derived from [æ] I chose the former. As for the vowels in "lamp" and "lap", I would posit that anyone who did not grow up in the Washington-to-Boston corridor does have the same vowel in those two words (except for the nasalization in "lamp"). Certainly they're the same vowel for me (I grew up in Texas but do not have a Texas accent). I don't know what region you mean by "MAE", but when you say that the vowel of "MAE" "can (n./vb.)" is the tense A, you have to realize that that is only true for regions affected by what Labov calls "Northern Cities Shift", which generally affects upstate NY, northern Ohio, Michigan, northern Indiana, northern Illinois, and Wisconsin. People from further west, like me, have only the lax A, so for us, the only difference between "lad" and "land" is the nasalization; both vowels are clearly more open than that of "led". Finally, whether the difference between lax A and tense A is slight or great (for people who have both) depends a lot on where the speaker grew up, and possibly socioeconomic background as well. My student (who was an undergrad at an Ivy League university and therefore presumably fairly well-off) had only a very slight difference in vowel height; her tense A was not diphthongized and not even significantly longer than her lax A. On the other hand, the (probably justified) stereotype of working-class NYC is that the name "Ann" is homophonous with "Ian", so for those speakers the vowel is clearly diphthongized, with a starting point probably even closer than [e]. --Angr 22:53, 19 Feb 2005 (UTC)

What you're not realizing is that tense A is a diphthong, but you are right that it can have some variation. I believe this variation to include [æ˔ə], [ɛə], [ɛ˔ə], and even [eə]. [æ˔ə] is very rare (I have rarely heard it used, except for maybe by one or two people). Around 90% of the population uses [ɛ˔ə] for tense A, and this same amount of people distinguish greatly between lax and tense A. In fact, a very small percent of the population, virtually none of whom are native speakers, only use lax A. The difference between "lad" and "land" is certainly not just nasalization. In fact, there is absolutely no nasalization in the word "land," as by no means does the back of the tongue ever reach the velum during pronunciation of this word. The word is transcribed as /lɛ˔ənd/, with a tense A, by virtually every native English speaker. Go to

http://www.geocities.com/jordanekay/voweldifferences.wav

to listen to the drastic difference between [æ] and [ɛ˔ə]. Perhaps you don't speak MAE (though if you don't distinguish between "marry," "merry," and "Mary," you do), but I don't understand how tense A sounds more open than [ɛ]. In fact, lax A, [ɛ], and tense A are spread out evenly in that order, shown here:

http://www.geocities.com/jordanekay/threevowels.wav

What is Modern American English? It is the form of English known as "accentless" English, meaning it contains none of the distinguishing characteristics of regional accents. I could go into greater detail, but you get the idea.

--Jordan 24.30.22.78

What I'm saying is tense A isn't always a diphthong; I know it isn't because I have heard nondiphthongal tense A's, for example my student's. Where do you get the 90% statistic for use of tense A? Is that supposed to be 90% of all speakers of American English? Or 90% of people in the northeastern quadrant of the country? Because I could believe the latter but not the former. There certainly is nasalization in "land", just as there is nasalization of every vowel before a nasal consonant. Nasalization does not mean the tongue is raised to touch the velum (which would result in a velar consonant); nasalization means the uvula is lowered to allow air to pass through the nose and mouth at the same time. I listened to voweldifferences.wav and agree that for that speaker (which I assume is you) there is a great difference between lax and tense A. But your tense A is a vowel that does not exist in my speech, nor in the speech of anyone I know who comes from the South, or anywhere west of the Mississippi. Unfortunately I couldn't hear anything when I tried to listen to threevowels.wav. But that doesn't matter; I believe you that you pronounce tense A with a closer starting point than . All I'm saying is that not everyone who has all three vowels produces them your way. --Angr 07:15, 20 Feb 2005 (UTC)

Your tense A in [1] (http://www.geocities.com/jordanekay/voweldifferences.wav) is very much regionally marked. Listen to some examples from speakers with regionally unmarked dialects and notice how similar the vowels are. [2] (http://www.nohat.net/nohat-lad-lamp.wav) [3] (http://www.nohat.net/jill-lamp-lap.wav) [4] (http://www.nohat.net/tom-lad-lamp.wav). The first is my own speech. The second and third were generated using concatenative speech synthesis systems. I won't get into the details here, but they are actual recordings of real people speaking, and the particular words given are guaranteed to be selected from single units. The voice talents used for those systems are selected by experts to be very neutral sounding. The second sample is a voice built on recordings of Jill Jacobs of jillsvoice.com (http://www.jillsvoice.com). The third sample is from a voice built on recordings of a singer named Tom whose last name (and website) escapes me at the moment. Listen to all three and you'll hear the only difference between the vowels is the nasalization. Nohat 09:00, 20 Feb 2005 (UTC)

As another example, I took snippets from Jill Jacobs' site of her saying "I'll ask" and "the answers" and concatenated them together at [5] (http://www.nohat.net/jill3.wav). Notice how the 'a' in ask has the same quality (except for nasalization) as the 'a' in answers. Note how it's not even slightly diphthongized. This is a classic example of standard American English. Nohat 09:21, 20 Feb 2005 (UTC)

I've just remembered Tom's last name is Glynn. His site is at tomglynn.com (http://www.tomglynn.com). He's the voice of for example the United Airlines automated flight status line at 1-800-824-6200. Another classic example of standard American English. No tense A distinction. No diphthongization. Nohat 09:29, 20 Feb 2005 (UTC)

Whoa, my tense A is definitely NOT regionally marked. Let's find speakers of American English who all use MAE and pronounce it just as I do.

Alyson Hannigan: "And one time, at band camp."- American Pie

http://www.geocities.com/jordanekay/bandcamp.wav

http://www.geocities.com/jordanekay/tenseA1.wav

http://www.geocities.com/jordanekay/tenseA2.wav

Jim Carey: "You slammed her! You dunked her doughnut! You gave her dog a snausage! You stuffed her like a Thanksgiving turkey!"- Liar Liar

http://www.geocities.com/jordanekay/slammed.wav

http://www.geocities.com/jordanekay/tenseA3.wav

Molly Ringwald: "I can't believe this."- Sixteen Candles

http://www.geocities.com/jordanekay/cant.wav

http://www.geocities.com/jordanekay/tenseA4.wav

But more importantly, there are absolutely no nasal vowels in the word "lamp." Listen as I say MAE "clamp" followed by Philadelphia "class." The only difference between the pronunciation two words (in the Philadelphia accent only) is their final consonants, "mp" and "s." Of course there is no nasalization in the word "class." Note that, in thie Philadelphia accent, the vowels are absolutely the same:

http://www.geocities.com/jordanekay/clampclass.wav

I certainly do not pronounce those words that way, and instead use tense A for the first and lax A for the second, as found in modern American English:

http://www.geocities.com/jordanekay/MAEclampclass.wav

If "the only difference between these two vowels were the nasalization," as you said, the words "lap" and "lamp" would sound like this:

http://www.geocities.com/jordanekay/laplamp1.wav (British English, no American says this)

or

http://www.geocities.com/jordanekay/laplamp2.wav (Strong midwestern American)

Finally, if these vowel sounds were the same, as you say, how could one distinguish between a Philadelphian's "pass" (tense A) and the MAE "pass" (lax A)?

http://www.geocities.com/jordanekay/passpass.wav

Clearly a difference is heard. Granted, there are many varieties of tense A, but most are diphthongized, and are used in modern American English before "n" and "m"; in Philadelphia English before "n," "m," "s," and "f," (all only at the end of a syllable), and in the words "mad," "bad," "glad," and "bath"; in varieties of New York (not upstate) English before "n," "m," "s," "f," "d," "g," and "sh" (all only at the end of a syllable); and in many Midwestern Accents (Chicago, northern Illinois), tense A replaces lax A altogether.

--Jordan

Jim Carey is Canadian, and his vowel of "slammed" sounded quite lax to me. Much laxer than your "lamp" at any rate. And I definitely hear nasalization in your pronunciation of "clamp". And I, for one, do have the same vowels in "lap" and "lamp", except for nasalization, and the vowel I have is lax and is more open than . I wish I had microphone to prove to you. My precise vowel quality is of course different from the British speaker's, but that doesn't change the fact that, like him, I have a single [æ] sound, and that sound is lax. The pronunciation of the speaker you call "strong midwestern" sounds like it's undergone Northern Cities Shift to me. Finally, I do not deny that the vowels are different in Philadelphia, New Jersey, and downstate NY! All I deny is that 90% of Americans use a tense A, either as their only realization of a single phoneme, or in contrast with a lax A. 90% of people in the Northeast and Inland North, yes; 90% of all Americans, no. I think if you took a microphone to any city west of the Mississippi (except possibly St. Louis which seems to be included in the Northern Cities Shift) and recorded people saying "mat, mass, met, mess" (preferably in carrier phrases like "Say ___ for me" so as not to get phrase-final lengthening) you would discover that there are lots and lots of Americans who (1) have the same vowel in "mat" and "mass" and (2) have a lax vowel in "mat" and "mass" that is more open (has a higher F1) than their vowels in "met" and "mess". Someday, Labov et al.'s Atlas of North American English will be published and we can look and see where [æ] is tensed and where it is lax. --Angr 23:47, 20 Feb 2005 (UTC)

There's a useful map at [6] (http://www.ling.upenn.edu/phono_atlas/maps/MapsIN/TelsurIN_eae.html), which only applies to people who do not contrast tense and lax A, and have but a single /æ/ phoneme. The blue dots represent people whose /æ/ is persistently tense, the variety you're calling "MAE". The green and yellow dots represent people whose /æ/ is persistently lax, the variety you said is spoken by only "a very small percent of the population, virtually none of whom are native speakers". I contend that to the west and south of the area on this map, we would find hardly any blue dots at all, and not many green ones either, but mostly yellow ones. --Angr 23:57, 20 Feb 2005 (UTC)

Your tense A is very regionally marked. If you talked like that in California and tried to tell people that you have a "standard" American accent, people would laugh at you. I know, because in my high school, people relocated from areas where there is a tense A/lax A distinction and promptly lost the distinction as well as any diphthongization due to social pressure. Also, your examples from movies aren't particularly valid, particularly the American Pie and Sixteen Candles examples, as both of those movies take place in an area with fairly advanced northern cities vowel shift, and as Angr said, Jim Carrey is Canadian. And even disregarding that, if you look at the specrograms of those vowels, you will see that there is very little diphthongization. On the other hand, the examples that I gave ARE examples of typical Standard American from people whose voices were specifically selected because their speech was neutral-sounding, and there is no distinction at all between tense and lax A. Tense A is just not a feature of Standard American English. It's a feature of your speech, but there's just not data out there that supports the claim that most Americans have a distinct diphthongized tense A. Your examples of your own voice that you recorded prove nothing other than the fact that you can produce different vowel sounds if you want to. Nohat 00:48, 21 Feb 2005 (UTC)

OK, I see where you're coming from. Tense A is a feature of Standard American English, however, no matter how close it may be to lax A in different accents, because almost every speaker (except those with Midwest accents) distinguishes between the A in "lap" and "lamp." Even Nohat's recording and the other two synthesized recordings make a distinction, using /æ/ for lax A ("lap") and /ɛə/ for tense A ("lamp"). (The two sound similar because [æ] and [ɛ] are only half as far apart as, say, /i/ and /e/ are.) To say that only my "regionally-marked accent" distinguishes is absurd, even if the distinction for other accents is very minor. However, tense A seems to be more commonly said /ɛ˔ə/ (the way I pronounce it) in modern American English. I am interested, though, in finding every possible way to pronounce every phoneme in English and conduct a large Internet survey, in which the responders indicate their age, all places of residence, residence of parents (from whom one may develop an accent), and the phonemes they use for different words, where every possible pronunciation in any accent is given as an audio recording.

--Jordan

First of all, I don't think it's useful to talk about "Standard American English". It's obvious that what Jordan considers standard is different from what Nohat and I consider standard. (Probably if we got into other examples, Nohat and I would disagree on what's "standard", too.) Except for people who are aware that they have regional accents (Southern, New York City, Boston, etc.), probably almost everyone assumes that their speech is "standard". Even the usual linguistic term "General American English" is misleading because there's so much variation in it. GAE includes accents with and without A tensing; GAE includes dialects with and without the whine-wine merger; GAE includes accents with and without the cot-caught merger. You say, "almost every speaker (except those with Midwest accents) distinguishes between the A in 'lap' and 'lamp.'" That statement is true provided you define "Midwest accent" to include everything west of the Mississippi as well as large areas of the South. But it is also true that nasalization can affect the acoustics of a vowel, which means the vowels of "lap" and "lamp" may have slightly different formants for that reason. Therefore when deciding whether or not someone has both tense and lax A it would be better to test them on pairs like "pat/pass", "lap/laugh". If those pairs are distinct, the person has both tense and lax A. If those pairs are the same, the person has only one kind of A. Then the question is, is that A tense or lax? To answer that, compare "pat/pass" with "bet/Bess" and see which vowel has the higher F1. If "pat/pass" has a higher F1, the speaker has a lax A (yellow and green dots on the map); if "bet/Bess" has a higher F1, the speaker has a tense A (blue dots on the map). If a speaker has the same vowel in "pat" and "pass" but a different vowel in "pan", then the difference is due to the effect of nasalization, not the effect of tensing. --Angr 06:44, 21 Feb 2005 (UTC)

I see we're still not coming to terms with this. I have realized however, that you were right in that vowels become nasalized before nasal consonants, kind of. In fact, there is a sort of nasal glide toward the end of the vowel as to prevent a gap between the oral sound of the vowel and the nasal sound of the consonant. However, many have taken this nasaliziation as the reason why the A's in "lack" and "land" sound so different, when in reality, they are different vowels altogether, for those that make the distinction. I will use the ~ nontheless to denote this nasal glide, for lack of a more appropriate symbol. I'm now forced to go through every one of your claims and refute them with phonetic evidence.

First of all, I don't think it's useful to talk about "Standard American English".<It's obvious that what Jordan considers standard is different from what Nohat and I consider standard. Even the usual linguistic term "General American English" is misleading because there's so much variation in it.

I am not describing Standard American English or General American English but rather Modern American English, the the most prevalent form of American English. As we speak, the numbers of speakers of MAE are increasing exponentially, while the numbers of speakers with accents are decreasing. This a result of many factors: one, MAE is the most "natural" way to speak English, as it contains the most optimum phonemes for the ease of movement of tongue and lips. For example, "land" used to be pronounced /læ~nd/, but in MAE, spoken by about 80% of the entire population of the United States of America (including myself, and, based on the arguments you have been providing, not inlcuding Nohat and Angr), it is pronounced /lɛ˔ə~nd/. Since the tongue position when pronouncing this word must be preserved until the d, a longer time than in words in which short a comes before a non-nasal, such as k or t, it is more taxing on the tongue and requires more muscles to keep the tongue in the fairly-open /æ/ position while the tip is curled up the alveolar ridge, than it does to curl it up while the tongue is in the /ɛ˔/ position, during which the tip is much closer to the ridge. The resulting /ə/ is the result of the creation of a two-step process of the curling of the tongue, even more optimal than a one-step process. This two-step process is relatively new, seen in the fact that it is used in MAE but not in the older forms of standard English spoken by Nohat and Angr, in which they have made clear they use a monophthongal vowel. Another reason for the advent of a modern American English is the invention of and importance placed on the television, which allows children learning to speak English to hear the language not only the way their parents and the people that have the accent associated with their hometown speak it, but the way many people from across the United States speak it. Virtually, the only way a child born in the year 2005 will develop a strong regional accent is if both parents have almost exact accents. If the parents have conflicting accents, the unsuredness of which phonemes to use for certain words leads the child to pronounce it the most natural way. Take, for example, a child born to two parents, one with a Philadelphia accent and the other with no accent at all (one close to MAE). The way this child learns such words as "water" is difficult, because the parent having the Philadelphia accent will pronounce it [wUdr] (a key feature of the accent for that word only) and the other will pronounce it [wQdr]. Hearing these conlflicting phonemes used, the child will choose the one that is used more often and the one he hears used by others and actors on TV: /Q/. Take another word, "banana." The Philadelphian pronounces this /bənænə/, with a lax A, while the other pronounces it /bənɛ˔ənə/, with a tense A. When faced with this conflict the child will choose the more natural, more commonly used tense A for this word. In both of these cases, the child has used the phonemes used in modern American English, and none of the father's Philadelphia accent is passed on to the child. The degrees to which this phenomenon occurs vary; sometimes children pick up many but not all features of the parents accents, and many times they do not pick up any features and develop a full MAE accent. The following adults, who use a modern American English accent, experienced this phenomenon as they learned to speak. Note how each pronounces each word almost exactly the same: [7] (http://classweb.gmu.edu/accent/english32.html), [8] (http://classweb.gmu.edu/accent/english59.html), [9] (http://classweb.gmu.edu/accent/english55.html), [10] (http://classweb.gmu.edu/accent/english60.html), [11] (http://classweb.gmu.edu/accent/english62.html), [12] (http://classweb.gmu.edu/accent/english63.html), [13] (http://classweb.gmu.edu/accent/english71.html), [14] (http://classweb.gmu.edu/accent/english74.html), [15] (http://classweb.gmu.edu/accent/english98.html), [16] (http://classweb.gmu.edu/accent/english96.html), [17] (http://classweb.gmu.edu/accent/english104.html), [18] (http://classweb.gmu.edu/accent/english50.html), [19] (http://classweb.gmu.edu/accent/english49.html). There are some slight variations in each of these, but they are incredibly minute and are most likely due to intonation rather than the phonemes themselve, which can account for phonemes, though usually the same, being slightly different.

GAE includes accents with and without A tensing; GAE includes dialects with and without the whine-wine merger; GAE includes accents with and without the cot-caught merger.

Correct, but MAE certainly includes the whine-wine merger, the cot-caught merger, and a severe lax A/tense A distinction.

You say, "almost every speaker (except those with Midwest accents) distinguishes between the A in 'lap' and 'lamp.'" That statement is true provided you define "Midwest accent" to include everything west of the Mississippi as well as large areas of the South. But it is also true that nasalization can affect the acoustics of a vowel, which means the vowels of "lap" and "lamp" may have slightly different formants for that reason.

Remember, it's simply a nasal glide toward the end of the vowel. This barely, if at all, affects the acoustics of the vowel. There is no audible difference between /An/ and /A[nasal glide]n/.

Therefore when deciding whether or not someone has both tense and lax A it would be better to test them on pairs like "pat/pass", "lap/laugh". If those pairs are distinct, the person has both tense and lax A.

Correct.

If those pairs are the same, the person has only one kind of A.

The most incorrect and ludicrous thing I've heard you say. These pairs are the same in MAE and that definitely uses two kinds of A's. The nasalization affects nothing. If the vowels in Philadelphia "pat/pass" are treated as different, then certainly are the vowels of "lack/land" (same as Philadelhpia "pat/ass") of MAE.

Then the question is, is that A tense or lax? To answer that, compare "pat/pass" with "bet/Bess" and see which vowel has the higher F1. If "pat/pass" has a higher F1, the speaker has a lax A (yellow and green dots on the map); if "bet/Bess" has a higher F1, the speaker has a tense A (blue dots on the map).

I agree.

<i>If a speaker has the same vowel in "pat" and "pass" but a different vowel in "pan", then the difference is due to the effect of nasalization, not the effect of tensing.

Again, nasalization makes no difference. If you establish that A) there is no nasalization in Philadelphia tense A "pass" and B) that is the same exact sound used in MAE "pan," it is clear that nasalization affects the vowel.

Please take what I say critically, as I have critically considered every point brought up in this discussion. This argument is simply for the veracity of the article, which I feel should be a true represenation of the phonemes used in American English.

--Jordan

Well, for all our disagreements, at least we agree on the point of the article!

• I am not describing Standard American English or General American English but rather Modern American English, the the most prevalent form of American English. As we speak, the numbers of speakers of MAE are increasing exponentially, while the numbers of speakers with accents are decreasing.

Okay, I thought you were using the term MAE interchangeably with what Nohat called SAE and I call GAE. I was mistaken, I'm sorry about that. But your term is misleading, because it implies that (virtually all) Americans nowadays talk that way. And what Nohat and I have been trying for the last four days to convince you of is that it is only Americans from a certain region of the country, namely the portion on the map I linked to above, plus probably New England, who talk that way.

• For example, "land" used to be pronounced /læ~nd/, but in MAE, spoken by about 80% of the entire population of the United States of America (including myself, and, based on the arguments you have been providing, not inlcuding Nohat and Angr), it is pronounced /lɛ˔ə~nd/.

Once again, where are you getting these statistics from? First you said 65% of adults and 80% of children speak MAE, then you said 90% of everyone, now you're saying 80% of everyone.

• Since the tongue position when pronouncing this word must be preserved until the d, a longer time than in words in which short a comes before a non-nasal, such as k or t, it is more taxing on the tongue and requires more muscles to keep the tongue in the fairly-open /æ/ position while the tip is curled up the alveolar ridge, than it does to curl it up while the tongue is in the /ɛ˔/ position, during which the tip is much closer to the ridge. The resulting /ə/ is the result of the creation of a two-step process of the curling of the tongue, even more optimal than a one-step process.

I'm not convinced by this articulatory explanation, but that doesn't matter. I don't deny the veracity of the sound change in question, and whether it is triggered by "articulatory ease", or by acoustics (nasalization causes the F1 of a vowel to fade away, making perception of vowel height difficult to recover; also, crosslinguistically vowel raising before nasal consonants is extremely common), or by sheer bloody-mindedness (as common a cause of sound change as anything) is irrelevant to the observation that the sound change has happened.

• Another reason for the advent of a modern American English is the invention of and importance placed on the television

That television can have a significant effect on linguistic performance is, I believe, no longer controversial. The number of Americanisms in British English has increased exponentially since American shows began to be shown in Britain, and the fact that far more people from the Netherlands and Sweden speak English well than people from Germany is usually attributed to the fact that in the Netherlands and Sweden, English-language shows are subtitled, while in Germany they're dubbed.

• Remember, it's simply a nasal glide toward the end of the vowel. This barely, if at all, affects the acoustics of the vowel. There is no audible difference between /An/ and /A[nasal glide]n/.

That's because no English speaker would ever produce a completely nonnasal [æ] before a nasal consonant. It's actually very difficult to do: the only way I can do it is to pretend I'm saying [æd] and then suddenly switch from [d] to [n] at the very instant my tongue touches my alveolar ridge. It sounds very odd and thoroughly un-English.

• • If those pairs are the same, the person has only one kind of A.
• The most incorrect and ludicrous thing I've heard you say. These pairs are the same in MAE and that definitely uses two kinds of A's. The nasalization affects nothing. If the vowels in Philadelphia "pat/pass" are treated as different, then certainly are the vowels of "lack/land" (same as Philadelhpia "pat/pass") of MAE.

I'm sorry, I should have been more explicit. I meant, if those pairs ("pat/pass", "lap/laugh") have the same vowel, the person has only one /æ/ phoneme. Because if the only environment where occurs is before a nasal consonant, and lax [æ] never occurs there, then the two are in complementary distribution and satisfy the test for being allophones of a single phoneme. This is true even if the exact same sound is a separate phoneme in Philadelphia!

--Angr 07:00, 23 Feb 2005 (UTC)

Well, I'm glad we agree on most things, like the influence of television on accent development. And yes, my "statistics" are quite arbitrary, and it will be a while before I finish my accent survey and administer it to a random sample of the United States population. But it is true that the nasalization of the short A in "land" is not what causes it to sound different than that of "lack." If it did, then how come there is no audible difference between the vowels in "cop" and "con," "met" and "men," and "keep" and "keen"? And the reason why I call it modern American English is that, as more and more children develop this accent every day, it will eventually be spoken throughout the country. Also, someone who speaks modern Modern English is never said to have an accent (except maybe facetiously by those with thick regional accents), while someone that does not use the MAE phonemes is said to have an accent. For this article, I do think that at some time it would be appropriate to include the phonemes used in MAE where they differ from older forms of standard English, such as /Q/ being used for "aw" instead of /O/. (I do agree that the broad transcription [O] can be used to denote any "aw" sound, however, but actual IPA symbols are extremely specific.)

--Jordan

• someone who speaks modern Modern English is never said to have an accent (except maybe facetiously by those with thick regional accents)

And by linguists, who acknowledge that everyone has an accent. As for using to represent the American vowel in "caught" (for those who still distinguish it from "cot"), J.C. Wells already does that in the Longman Pronunciation Dictionary, though in his case it's probably to emphasize how much more open the vowel is in American English than in British English, for which he uses traditional . --Angr 06:26, 24 Feb 2005 (UTC)

I was under the impression that the cot-caught merger described that the vowels are very similar but not exactly the same. The only regional accents I can think of that do use the same phonemes are those that use /Q/ for both, like Chicago, Boston, and Pittsburgh, and some other Midwestern accents that use /a/ for both. MAE uses /A/ and /Q/, which sound similar; very rarely in MAE does "caught" rhyme exactly with "cot." Do you know of any accents that use /A/ for both "ah" and "aw"?

--Jordan

Sure, California. In fact, probably the entire West. But I thought Chicago still kept them distinct, as indicated by the map at [20] (http://www.ling.upenn.edu/phono_atlas/maps/Map1.html), where blue dots are those who keep them distinct and red dots are those who merge them. (At that page they use /o/ to mean and /oh/ to mean for some weird historical reason.) --Angr 04:55, 25 Feb 2005 (UTC)

Yes, the merger is definitely total for most native Californians, as well as a for a large swath of the rest of the country, and there are even substantial numbers of speakers in areas that supposedly maintain the contrast for whom they merger is total, especially among younger people. I did a semi-scientific survey of around 50 people when I was in college, and even found people with the merger who grew up in and around New York City. Nevertheless, the distinction remains alive and well in many areas, but for me, at least, it didn't even occur to me that cot and caught might NOT be homonyms until I studied linguistics formally. I never really thought about it very hard before studying linguistics, but I was always baffled as to why dictionaries had two different symbols for what was obviously (to me) the same sound. Nohat 06:15, 25 Feb 2005 (UTC)

Jordan—you seem to be confusing co-articulatory effects with articulatory targets. A co-articulatory effect is when the physical mechanics of articulating sounds change the way certain sounds are produced. For example, nasalization of vowels in English is a co-articulatory effect. What happens is that the production of a vowel followed by [n] involves three articulatory gestures. First, the tongue is placed for production of the vowel. Next, the velum is lowered to permit airflow through the nasal cavity. Finally, the tongue tip moves from its vowel production position to the closure for production of the nasal, which in the case of [n], involves moving the tongue tip to the alveolar ridge. In the idealized world of the IPA transcription [æn], the second and third steps occur at the same time. There are some languages that distinguish nasal vowels before nasal consonants (i.e. [æn] and [æ̃n] could be a minimal pair), so in these languages the timing of the lowering of the velum is important, as it carries a functional load in distinguishing words. In English, there is no distinction, so it is not important to delay lowering of the velum until the same time as the closure for the nasal consonant is made. So what happens is the velum gets lowered earlier to help distinguish between [n] and [d], which is what [n] would sound like if the velum weren't lowered in time. In fact, the distinction between [n] and [d] carries a significant functional load in English, so English speakers have been conditioned to always lower the velum early for [n] so as to ensure [n] and [d] are never confused. This explains the nasalization of consonants; it does not, however, mean that English has nasal vowels, it just means that vowels are nasalized in English before consonants as a co-articulatory effect. An articulatory target is the intended movement and positioning of the articulators. For speakers like who do not have tense A, the articulatory target for position of the tongue for production of [æ] is invariant with respect to following nasals. This means that the brain "program" that moves the tongue into place for the production of [æ] does not depend on whether the following consonant is nasal. The perception of raising is due to the effect on the acoustics of nasalization. The actual position of the tongue is not in fact higher. Similarly, the perception of a schwa-like glide between [æ̃] and [n] is due to the movement of the tongue to articulate the [n]. In order to place the tongue tip on the alveolar ridge for production of [n], the whole body of the tongue has to move up out of its low position for the production of [æ]. The same effect can be heard without nasals, for example it will be heard in ad but not in ab, which doesn't have this effect, because the articulation of [b] only involves movement of the lips.

In contrast, dialects which have tense A have a different articulatory target for the production of /æ/ when it's tense. This means that the brain "program" for the production of [æ] is different when [æ] occurs before nasals. This is what is meant by allophone: a different articulatory target without a change in functional load (or phonemic status). In other words the sound is produced differently, but means the same thing. Another example in English would be light and dark [l]. There are no minimal pairs—the sounds occur in complementary distribution—but the two sounds are produced with very different positioning of the tongue, which is clearly not just a coarticulatory effect. It is perfectly possible to produce the word pal with a light [l]; it just doesn't sound English. In American English, we can divide the dialects into three groups based on the production of [æ] before [n] and [m]: (a) dialects where [æ] is raised before [n] and [m] purely as a co-articulatory effect, but the articulatory target for [æ] is the same before nasals and non-nasals; (b) dialects where [æ] is raised fully to [ɛə] before nasals as an allophonic variation, where the articulatory target for [æ] is different before nasals than non-nasals; (c) dialects where [æ] and [ɛə] are separate phonemes before [n] and [m], with different degrees of raising depending the phoneme. However, the presence of a co-articulatory effect is not the same thing as a dialect which has a different articulatory target for the production of a sound. You seem to be pointing to the co-articulatory raising in dialects in group (a) as evidence for the dialects of group (b) being the standard. What is likely is that, as a speaker of a dialect in group (b) or (c), your ear is attuned to hearing raised [æ] before [n], and so interpret co-articulatory raised [æ] as being the same as your own different-articulatory-target raised [æ], when in fact it is not.

Note that this discussion does not consider the status of [æŋ], which doesn't occur in many dialects which are in the first condition described above. For example, in my dialect, the articulatory target of the vowel of bank is the same as that for bake, and not that for back. But we're only considering [æ] before [n] and [m] here, where it can clearly be contrasted with [ɛ] and [eɪ] (can/Ken/cane, cam/chem/came); no such minimal sets exist for the low front vowels before [ŋ].

I hope that this long-winded explanation explains the situation adequately.Nohat 07:39, 25 Feb 2005 (UTC)

Aha, thank you for that explanation; I agree with almost all of it. However, I am certain that very few, if any, dialects of American English have only co-articulatory effects on [æ] before [n] and [m]; almost all have at least some different-articulatory target effects, even to the smallest to degree as raising the phoneme, or start "program" of the brain as you have explained, to at least [æ˔] instead of [æ]. In accents that only have co-articulatory effects for [æ] nasals (such as many dialects of British English), the words "lap" and "lamp" would sound like [21] (http://www.geocities.com/jordanekay/laplamp1.wav). Thus I am not mistaking the co-articulatory effects of tense A found in all English dialects for the different-articulatory-target effects (is there a better term for this?) found in MAE. Plus, the co-articulatory effect on /A/ in the words "con" and "cot" certainly does not make those two vowels sound at all different to me (besides, of course, the final nasal glide in "con").

Also, modern American English does include a final schwa glide at the end of tense A, which is why the vowel is transcribed as [ɛ˔ə]. This is not just the moving of the tongue so the tip can reach the alveolar ridge; but an actual glide. This way the glide is also heard in "am" as well as "an." If there were no glide, the word "Ann" would sound like [22] (http://www.geocities.com/jordanekay/noglide.wav) [ɛ˔~n] instead of [23] (http://www.geocities.com/jordanekay/glide.wav) [ɛ˔ə~n], it's MAE pronunciation.

Finally, I am still not convinced that Californians, or any dialect, uses /A/ for both "cot" and "caught." MAE certainly differentiates between the two, using /A/ and /Q/, heard in [24] (http://www.geocities.com/jordanekay/cotcaught.wav). Nohat, you say that you don't differentiate between the two; could you provide a sample of your voice (or any other voice with the cot-caught merger) pronouncing the sentence "The rabbit I caught lept into the cot"? I need aural evidence of both these words using the vowel /A/; the most merged I've ever heard them been, while still retaining a "back" pronunciation, is /A/ and /Q/. Thanks.

--Jordan

Sorry for the delay in reply, but the person who was going to record this for me properly was out sick the past two days. The following recordings are of my voice:

What a doll Dawn Dawley and Maude Lodz are. [25] (http://www.nohat.net/audio/c1.mp3)

You naugty ox—the gaudy hawk body you tried to hock is a knotty bawdy box. [26] (http://www.nohat.net/audio/c2.mp3)

Paul pawned Dolly's ox pond that was chock full of awk chalk for that awful clod God. [27] (http://www.nohat.net/audio/c3.mp3)

Fawley's folly was his stock yawn that awed yon Don but not odd Sean. [28] (http://www.nohat.net/audio/c4.mp3)

You'll note that there is no distinction between any of the A/O pairs. Can you tell which is which in the following? naughty-knotty (http://www.nohat.net/audio/naughty-knotty.wav) pawned-pond (http://www.nohat.net/audio/pawned-pond.wav). I promise the cot-caught merger is real and complete for the vast majority of native Californian English speakers. Nohat 09:22, 3 Mar 2005 (UTC)

Ah, thanks. So it is indeed real, but it's not a feature of modern American English. I had yet to hear someone use /A/ for both vowels until your clip. The question is, which is prevalent and will ultimately replace the other: /A/ and /A/ or /A/ and /Q/? Based on from what I've heard listening to speakers of all ages from every region, I'll place my bet on /A/ and /Q/. A cot-caught semi-merger, I presume?

--Jordan

One thing that's particularly interesting about this merger is the number of people who can't hear the difference they make. I read somewhere (I don't remember where now) about an experiment where people produced minimal pairs like "cot"/"caught" or "Don"/"dawn". Phonetic analysis of the vowels revealed there was a distinction, yet when the subjects were played back the recordings of their own voices in random order, they couldn't tell which was which. So they can make the distinction, but they can't hear it. --Angr 08:09, 4 Mar 2005 (UTC)

Can we drop SAMPA

With the addition of Australian versions, this article is getting to look a bit cluttered. Would this be the time to drop the SAMPA codings from this article - particularly now that we've got the IPA template, so that even those using Internet Explorer can read the characters? rossb 08:12, 28 Feb 2005 (UTC)

Sounds good to me. Be bold, Ross. --Angr 09:22, 28 Feb 2005 (UTC)

Indeed, I agree with everything Ross has said. SAMPA has slowly been removed or replaced by IPA in most other articles so why not here too. – AxSkov 09:55, 28 Feb 2005 (UTC)

OK, I've removed the SAMPA. rossb 21:57, 28 Feb 2005 (UTC)

Good. However, I've in put a reference to SAMPA in "See also" so it isn't lost completely. Thincat 09:54, 1 Mar 2005 (UTC)

Affricates

Since are not phonemically affricates, and are neither phonemically nor phonetically affricates, do they really belong in the section on affricates? Or do we want to add in words like leapfrog but point out that phonemically, it's a sequence of and and not an affricate? Isn't this page supposed to be about how to use the IPA for English, not a discussion of English phonology? --Angr 06:55, 10 Mar 2005 (UTC)

We have two options: include those non-phonemic affricates, or delete all of the non-phonemic sounds listed in this article, such as (dark l) and (the "U" sound). I believe the former is better because we discuss not phonemes but actual pronunciations here to explain IPA. Phoneme notations don't have to match IPA exactly (you can use for ). In addition, the affricate and the stop-fricative sequence are different in English because the former never has a syllable boundary while the latter always does. I can't find a good example, so suppose there is a word boots-trap. It is pronounced differently from boot-strap, even though they have the same phonemes. It's clear in leapfrog is not an affricate because there is a syllable boundary between and .

and ( and ) are really affricates, which is clear for non-English speakers like me. If you pronounce without the affricate, it would sound like to rain, not train. English speakers often describe them as "chr" and "jr" to explain their actual sounds. Could you please revert the deletion? However, it is all right if you move information of this article (including the affricates you have deleted) to English phonology, which doesn't have much information now. - TAKASUGI Shinji 10:46, 10 Mar 2005 (UTC)

I didn't delete the nonphonemic affricates, Nohat did. But I'm in favor of the deletion (and in favor of deleting as well), on the grounds that this page is supposed be about the IPA symbols used in English, and these nonphonemic entities don't have separate symbols, just sequences of two symbols. We should keep , though, because it's a separate symbol that nonlinguists might encounter and want to know what it means. As for whether and are phonetically affricates, I concede that in some accents (but certainly not all) there is affrication of [t] and [d] before . But this just means that there are dialects where (or something close to them) are allophone of /t, d/ occuring before , not that the is itself part of the affricate. --Angr 12:39, 10 Mar 2005 (UTC)

I concede that in some accents (but certainly not all) there is affrication of [t] and [d] before [ɹ].

That is bull. Find me one speech example in which someone who is linguistically unware does not pronounce it "chr" and "jr" and, well, I'd probably say that person was born some time in the early 1900s or is not a native speaker.

--Jordan

What about...

What about the sound in air, bear, spam, ail. I'm not sure if it is or in AmE, but it's not on here. Specifically I was wondering how to spell Charon in IPA: ?

I would think that Charon would be pronounced with an initial /k/, as this is the convention for words of Greek origin. rossb 19:33, 15 Mar 2005 (UTC)

The mythological figure Charon is pronounced or , depending on your accent, in American English. The vowel is that of Mary and bared listed under R-colored vowels on this page. According to Charon (moon), Pluto's moon was pronounced with intial by its discoverer. --Angr 19:42, 15 Mar 2005 (UTC)

But the sound in spam or ail is the same, but not r-coloured. Like a shortened form of .

The vowels of spam and ail are not the same. The vowel of spam is a slightly raised [æ] and the vowel of ail is a slightly centralized variant of the diphthong [eɪ]. The non-R part of the vowel of air is a slightly raised [ɛ]. Note how no dictionary uses the same symbols for the vowel of spam and the vowel of ail. Nohat 06:03, 16 Mar 2005 (UTC)

The vowel of spam is a slightly raised [æ]

Slightly my ass. It's way past [ɛ]. Face it, lax A and tense A are two completely different sounding vowels. The vowel in spam sounds nothing like that of "spat." Oh, so the dictionary uses the same vowel, big whoop. When will people realize that the American English of 50 years ago is different from that we speak today?

Not everyone has the same American accent you do. There are still millions of Americans west of the Mississippi for whom [æ] before a nasal is still only slightly higher than [æ] before a nonnasal, and still well lower than . --Angr 06:10, 24 Mar 2005 (UTC)

I would like auditory evidence of such a pronunciation. A lax A lower than exceedingly rare and is confined only to such regional dialects you mentioned, certainly not a feature of standard American English.

And yes, in standard American English, the vowels of "pail" and "spam" (without the /ə/ glide of "spam") are, indeed, exactly the same. Exact phonetic transcriptions of these words are spoken in standard American English are as follows: /pʰɛ˔l̴/ and /spɛ˔ə~m/. Maybe not how people with regional accents say them, but it is certainly the way people without a regional accent pronounce the words.

Jordan, we've been through all of this elsewhere. Having a prenasal allophone of /æ/ that is so tense it's higher than is a marker of a regional accent, namely the Inland North and the Northeast. In the rest of the country, and in General American, if /æ/ is raised at all before nasals, it's still lower than . If you want auditory evidence, go anyplace west of the Mississippi or south of the Ohio and listen to the way the locals talk. --Angr 23:16, 24 Mar 2005 (UTC)

Syllabic n or l

My dictionary (Collins) uses a superscript schwa to indicate syllabic n or l, rather than the dot notation. Which is standard? Are both? -- Tarquin 11:04, 25 Mar 2005 (UTC)

The vertical line under the letter is standard IPA. The raised schwa is common in dictionaries because it's easier to typeset (no worrying about getting the little line directly under the l) and easier to see. Some people don't do either, and just transliterate a word like buckle as , which is unambiguous. But the problem with doing that is that you then can't distinguish between the two-syllable and the three-syllable pronunciations of words like settler. --Angr 12:31, 25 Mar 2005 (UTC)

ʊə sounds

The chart entry for ʊə (near the bottom) says that in British English it is cure, poor and jury but those sounds aren't the same at all. Cure and jury are more cyore (or cyurh) and jyurie whereas poor is the same as pore

Which sound is the correct one for the IPA symbol? From the Row below which is Australian English, I would guess that it is the poor sound --MyNameIsClare 12:33 14 Apr 2005 UTC

It's true that many British speakers use in cure and poor, so that poor is homophonous with pore and cure rhymes with both. The symbol represents the sound used by those speakers for whom poor is still distinct from pour and for whom cure rhymes with poor but not with pore. The sound spelled is a diphthong that begins with a vowel like that of foot, or partway between that of foot and that of goose, and then moves to the schwa sound of sofa; it sounds a bit like "oo-a". Some people in England do still pronounce poor and cure with that vowel, but it's losing ground. --Angr/_comhrá 12:59, 14 Apr 2005 (UTC)

Oh okay, I see what you mean. Yes some people do say cure and poor like that, but not many. It doesn't seem helpful to have those examples in the table without mentioning that as I think it would confuse people who pronounce them in the more common way. And I don't think I have ever heard anyone say jury like that! --MyNameIsClare 13:10, 14 Apr 2005 (UTC)

I must be another "just weird" person, but I am an American (Californian) speaker who pronounces poor , cure , and jury . Anyone care to explain this away? Is this dialectal or idiosyncratic? Also, my pronunciation of the vowels in father and car is more like [ɐ]. --PlatypeanArchcow 01:35, 27 Apr 2005 (UTC)

It can't be explained away. The ugly truth is that there is virtually no word in English that consistently has for rhotic and for nonrhotic speakers. I predict that in a century or two there won't be an accent of English that even has that phoneme anymore. --Angr/_comhrá 06:11, 27 Apr 2005 (UTC)

I'm a British/English Yorkshireman, and have spoken the language for over 50 years. Most of this article, to me, at least, reads as clear as Chinese in an Urdu dialect. The article needs copyediting and whittling down to a basic page, before it becomes an encyclopaedia in it's own right.

I know that this anon comment is from long ago, but I thought I would respond: This is a technical article about the phonetics of English. The fact that you've spoken English for 50 years doesn't necessarily make you able to read and fully understand a technical article about phonetics any more than the fact that you've lived in the physical world for 50 years should make you able to read and fully understand technical articles about physics. The page is pretty tight as it is—everything on this page is explained and can be read about more fully by following links. If it were whittled down any more it remove useful information. Nohat 05:32, 13 Jun 2005 (UTC)

Pronunciation guides in Wikipedia articles

Is there a guide for how best to provide pronunciation explanations at the top of articles and that also contains the Unicode values ready for copying?

I remember seeing a discussion about what best to use for pronunciation guided, but i can't find it anymore. The majority argued for IPA, though. If someone remembers, pleaese let me know.

Now my concrete question: The article wort says: "(pronounced wert)". I wanted to convert this to IPA and I would assume this maps to this page's example "bird: AmE , BrE & AuE ", but i have the following problems:

• How to best write this in an article. Maybe like this: "wort: [[AmE]] , [[BrE]] & [[AuE]] "? This looks very clumsy to me.
• Where should BrE and AmE point? To British English and American English or to Received Pronunciation and General American?
• Can we avoid splitting up the pronunciation? and are just regional variations of the same phoneme, so i think Wikipedia should rather refer to the phoneme rather than the exact pronounciation. Is there a way to refer to the phoneme?

Thanks! — Sebastian (talk) 05:48, 2005 May 15 (UTC)

I think the policy is at Wikipedia:Manual of Style (pronunciation). It is Wikipedia policy to use IPA. I agree that pronunciation guides in nonlinguistics articles should be kept as simple as possible. For wort I'd probably use something like as a compromise between rhotic and nonrhotic pronunciations. If you do want to split them up, I'd write something like: "RP /General American ", because there isn't only one British or one American pronunciation. I wouldn't worry too much about Australian pronunciation unless it's an Australia-specific topic, because at the phonemic level Australian English is so close to RP (the differences are mostly in the distribution of the broad A; otherwise it's mostly a matter of the realization of the phonemes). --Angr/_comhrá 10:39, 15 May 2005 (UTC)

What?? As far as I know, Australian English pronunciations are very different from that of RP. -- Mark 14:26, 15 May 2005 (UTC)

Yes, the surface pronunciations are very different. But at the phonemic level there's very little difference; the biggest difference is that RP has the "broad A" in many words where Australian has the "flat A": advantage, chance, demand, example, plant have /æ/ for most Australians but in RP. Also Australian has the weak vowel merger, so that abbot and rabbit rhyme, but they don't in RP. But Australian agrees with RP in merging /ɑr/ with /ɑː/ so that farther and father are homophones, in merging /or/, /ɔr/, and /ɔː/ so that hoarse/horse, source/sauce are homophones, in having /hw/ -> /w/, etc. It's how the phonemes are pronounced on the surface that's the big difference between the two, not how the phonemes are distributed in words. --Angr/_comhrá 14:46, 15 May 2005 (UTC)

Thank you for your replies and for the link to Wikipedia:Manual of Style (pronunciation). Unfortunately, that page doesn't say anything about splitting up pronunciations. I would like to reduce the need for parentheses or splits altogether. How about if we decide for just one of either or and refer to this page? Here, they would find (instead of the current definition lists) tables like this:

Phoneme	example(s)	AmE	RP	AuE	Notes
	bird	align="center"	align="center" colspan="2"
	beard	align="center"	align="center"	align="center"

I'm not aware of any ambiguity that would be introduced by this pair, and i'm hopeful that applying this for all phonemes generally doesn't create much ambiguity, so taht we could drastically reduce the need for parentheses or splits. — Sebastian (talk) 04:01, 2005 May 16 (UTC)

In principle that would be great, but good luck getting people to abide by it. If you use and in pronunciation guides for words relating to British or Australian topics, you'll immediately be reverted and probably called an American cultural imperialist. If you use them in words on country-neutral topics (like wort), someone will probably add the tag "American English" to it and then include a non-rhotic equivalent. Regardless of the useful chart you've drawn above. I think you're better off leaving well enough alone. --Angr/_comhrá 04:33, 16 May 2005 (UTC)

Can I be radical and argue against the IPA or is that an absolutely hard-and-fast limitation? Telling people that 'wort' is pronounced \WERT\ is much more helpful and unlikely to cause dispute than saying it's pronounced /w2:t/ (NZ) or /wr=t/ or /w@:t/ or whatever is the current habbit for whatever region today. Also, it provides no argument for 'cultural imperialism'; if there's two different pronunciations (as with dance) at the superphonemic level, you'd need to list the two, but the three pronunciations I've mentioned above are superphonemically the same (even if phonemically they imply different things for rhyme or rhythm or something).

(Also, there's a bunch of growing differences between AuE and RP, and even varieties of AuE. Phonemically RP I believe treats URE (pure) and YAW (yore) the same, but with one or two exception like 'sure' and 'your', AuE maintains a distinction. Also, 'celery' and 'salary' are homophones as /sæl@ri/ (for me a Victorian), as are 'due' and 'Jew' as /dZu:/; on a higher level the only way to read \DEW\ or \SEL-e-ry\ is the same as the only way to read \JEW\ and \SAL-e-ry\, but it's no more an issue of phonemics than J and G both spell the same sound in 'judge'. [The fact that I can't hear the diff. b/n 'celery' and 'salary' w/out trying real hard possibly suggests that's at a different level.])

However, that's not to say that the IPA's never useful here. Frex, if the actual focus is a distinction, for instance, the IPA'd be useful.

I added spelling-pronunciations like you suggest to the tables of the natural satellites of the solar system, as well as to the asteroids, with keys like ar as in car, etc. It's worked pretty well. Someone has added IPA to some of the entries, but I think the scope of the effort put them off. The only complaints I've had are that digraphs are inherently inappropriate because they represent single sounds, which therefore should be written with single letters. A few squirts of ink by the several of us working on this and those people quieted down. The strongest complaint was that ah for the a of father and oh for the o of bone were incorrect because those vowels have no aitch sound. The person insisted that I was pushing American pronunciation, and that in Britain ah means the sound of German ach. We settled on oe for bone and aa for father. There's a pronunciation key at Asteroid pronunciation key, if you're interested. I'm not claiming that's any kind of standard, but it's worked so far, and it would be nice if we all used a single system. Also, if you can give an Aussie input, that might improve the setup. I'm sure I missed some low back vowels in there somewhere! kwami 03:42, 2005 May 26 (UTC)