more on vowels

Being too lazy to make myself a sandwich today, I went out for a so-called Reuben. There in the bagel shop what looked like a mother and daughter were having a disagreement. The daughter was pronouncing castle like I would--with the same 'a' sound as in bat. The mother took her to task for this--saying it is pronounced with an 'ah' sound (like in father).

The daughter's response was one that an American (at least not outside eastern New England) would never say: "It's not 'cah-stle' because it doesn't have an 'r' in it!"

(The mother went on to take the daughter to task for trying to be someone she isn't. Unfortunately, I missed the relevant bit when she said "You were born in Brighton, not [inaudible]!" Could have been someplace in the US, could have been someplace in the UK--since the 'ah' vowel-before-/s/ 'rule' is a particularly southern/'standard' English kind of thing.)

Americans who speak dialects with post-vocalic (i.e. 'after vowel') /r/s tend to think of ('standard') BrE accents as 'leaving out the /r/s'. But the daughter's response here gives an indication that that's not how BrE speakers perceive their dialects. Instead, they perceive the /r/ as being part of the vowel sound. (Really, the /r/ and the vowel preceding it are pretty much a single merged-together sound in any dialect--that's the nature of /r/. It's also one of the things that bad impersonations of AmE accents tend to get wrong--e.g. saying farm as fah-rrm. I had an old SAfE-speaking boss whose impersonation of AmE accents was spot-on except for that detail.)

And all this goes back to the previous post from today, about tot and that it's pronounced differently from tart. AmE speakers may think of a ('standard') BrE pronunciation of tart as 'not having an /r/'. But that's not necessarily how BrE speakers are perceiving (or even pronouncing) it.


  1. There is a limerick I cannot see from where I sit in Kinglsey Amis's anthology of light verse, by Robert Conquest, which plays entirely on this vowel sound, about the surrealist peer Lord Garsole, who had himsef mailed in a parcel addressed just to "The castle, Garsole" .. but the rhyme you are waiting for never arrives. I think it is the Oxford book of Light Verse.

  2. I've found that "khaki" and "car key" make up a good test case. My (American) wife insists that she can't hear any difference between my pronunciation of them, whereas I (English) obviously can - primarily in the length of the 'ah' sound. Even exaggerated, she doesn't hear any difference; presumably it's only the sounded /r/ that she hears, or not, and the vowel length doesn't figure much.

  3. Lynn (or anyone else originally from the NE U.S.), do New-Englanders hear a difference between vowels with 'r's and those without.

    Regarding a different vowel sound, I know that as a Western U.S. speaker, I don't hear a difference between tot and taught. However, a friend from New York insists there is.

  4. Can the R droppers hear dropped Rs at the end of words? For example, do you pronounce 'tuna' and 'tuner' the same, or is there a subtle difference in vowel length? Similarly, 'tuba' and 'tuber', etc.

    What about the added dropped R? In the sentences "The tuba/tuber is yellow", the R is dropped (to my American ears) because it always is, and then it's added back in because it's between two vowels. Do the two sentences sound the same to the R-less?

  5. I had a similar vowel situation not long ago, regarding the correct pronunciation of the name "Tanya" (referring to a specific person from London, not the name in general). I asked a UK colleague who knows her very well whether her name should be pronounced Tanya [using a vowel as in "bat"] or Tanya [using a vowel as in "father"].

    He was fairly puzzled, eventually deciding it should be the former "because it isn't spelled with an R". My first thought was that that didn't make any sense at all; I didn't give him a choice of "TARN-YA".

    Of course, now I realize I had given him the extremely difficult challenge of trying to map my own (US/upper midwest) vowels onto his own, plus then trying to guess which of them she would prefer (turned out to be the latter).

  6. I've been meaning to write about khaki/car key, since I can't tell them apart when BrE speakers say them. (In AmE, khaki has the bat vowel, which makes BrE speakers giggle, since it sounds like cacky (='poopy').

    I can't answer questions re how r-less Americans perceive the sounds, since I'm entirely r-ful. But I can say that Southern and NYC r-lessness sounds different from BrE r-lessness in terms of vowel quality.

    The number of puns based on -er/final a (not thinking of one right now!) hints that at the ends of words the sounds are perceived as being fairly similar.

    Like Dunce, I've been thrown for a loop by the BrE pronunciation of Tanya--and also of a lot of other 'foreign' words that have acquired the bat vowel in BrE--like pasta. Probably worth another post at some point!

  7. It could have been someplace like 'Blaydon', if she was trying to be alliterative.

    The a sound is one of the 'caricature' differences between Northern and Southern accents (well at least from my southern BrE perspective)

    castle and bath are the most common example that I have come across. Northerners say them with the 'a' sound from bat - us southerners tend to say them with a short 'ar' sound.

    Ooh - this is when I wish I knew all those fancy pronunciation codes ....

  8. lynneguist said...

    > The number of puns based on -er/final a (not
    > thinking of one right now!) hints that at the ends
    > of words the sounds are perceived as being fairly
    > similar.

    A perfect example, although it is somewhat complicated by ignoring the toona/tyuna difference:

    > Like Dunce, I've been thrown for a loop by the
    > BrE pronunciation of Tanya--and also of a lot of
    > other 'foreign' words that have acquired the bat > vowel in BrE--like pasta.

    Um, re: 'acquired'. I might be wrong here, but isn't the typical BrE (at least where I come from) pasta (with the bat vowel) closer to the Italian than the AmE pahsta?

  9. Thanks for the cartoon link, Paul.

    Italian doesn't have the /ae/ vowel in bat, it has the /a/ vowel in father (see here. So, no, the BrE version is not more like the Italian. And the BrE version of Mexican food words, like taco and (in some people's pronunciation) fajita, also often substitute /ae/ for /a/--getting further away from the source language. I have started a few posts on this issue--but it's fairly complicated, so they might not see the light of Blogger for a while...

  10. Complications compound when you consider changes over time. Listen to dialogue in American films of the 30s, and compare with today. Vowel sounds used to be more clipped, closer to BrE. Today there's more drawling.

    Dot Com has become Dart Calm. Suicide bombers are now suicide balmers.

  11. I actually did a little research on the 'added' and 'omitted' r's during my brief foray into academic linguistics. It was of some interest to me since, being from Massachusetts, I often noted the assumption of non natives that Mass. residents spoke like Bostonians (If not 'Bert and I' clones).

    I gather Lynn would be a better source than I, but I believe that the Connecticut river (just on the western side of which I was raised) is a basic border of this trait, and that people only a couple of towns east have noticeably closer to Boston pronunciation. This has become less pronounced (so to speak (ditto)) with the introduction of bridges, cars and mass communication, but it still holds. West of the connecticut (at least here in Western MA, we are closer to Albany, NY than Boston in speech habits.

    I gather that the trait came from fashionable a fashionable London affectation that was also preserved in the Richmond, VA area. If anyone has seen (heard) the new Ken Burns effort "The War," the pronunciation by a Mobile, AL (nowhere near VA) lady of 'war,' which she understandably says quite a bit, sounds to my ear like it was impossible to utter without effort. It was something like /wah'ah/, whereas I would rhyme it with 'wore' (more or less).

  12. As a Washington-State-raised now-Californian who's spent much time in South Asia where RP is the prestige pronunication:

    * RP versions of "khaki" and "car key" sound different in my head. (But then I speak Spanish and can hear the length difference between two successive /a/s in "va a ver" ["is going to see"] and three successive /a/s in "va a haber" ["is going to be"].)

    * I grew up pronouncing "tot" and "taught" the same and I suspect I still do. But I can hear the distinction in an East Coast (or UK) accent and can reproduce it if I try. Especially if I imagine an overdone New Jersey accent.

    * The number of puns based on -er/final a: There's a children's program in the US called "Dora the Explorer". Most of the country thinks those words don't rhyme, though apparently someone involved in its creation think they do.

    * the 'ah' vowel-before-/s/ 'rule': Lynne, is this rule invariant? I recently corrected a non-English speaker's pronunication of NASA as /nasa/, saying that /næsa/ was correct. Were they just using a non-American pronunciation for an American organization?

  13. Here in Southeastern Mass., you will usually here the r-reversals (i.e, dropping word- and syllable-final r's that most Americans pronounce, and adding an r sound to the end of words with final short 'a'), but the Anglicized pronounciation of the sound most Americans normally sound as 'ae'=bat, well, that is fading here, increasingly restricted to locations along the coast up to Maine, and to people who are older. I am not sure why that is, but tv must have a great deal to do with it. As for those 1930s American movies and newsreels, it is true that the prestige stage dialect used by American actors then was far closer to RP English than anything any American actors or news readers use today, but that was, IIRC, more or less because it was an artificialized Anglo-American stage dialect designed to be some sort of via media between the two national norms in the first place, and Americans would probably never go for that today anyhow (can one imagine Rambo speaking with such a dialect?)

  14. The number of puns based on -er/final a (not thinking of one right now!) hints that at the ends of words the sounds are perceived as being fairly similar.

    One of my favo(u)rites is the Led Zeppelin song "D'yer Mak'er". For many years I cluelessly pronounced it (and interpreted it) as something more like "do your maker", with no idea that it made reference to the old joke:

    A: My wife went to the West Indies.
    B: Jamaica?
    A: No, it was her idea.

  15. Anon, the 'ah'-before-/s/ rule is undoubtedly more subtle than I've painted it as here (otherwise people wouldn't be saying paesta instead of pahsta!) Names, like NASA, are in particular resistant to rules that affect other parts of language.

  16. As an aside regarding vowels. My girlfriend (raised all over the US) will always try to correct my pronuncuation of the letter A, but only in certain instances.
    For example, when I say the name Aaron, it does not get confused with the name Erin, as the A has a much rounder sound to it. The same thing happens when saying the name Carrie as opposed to Carey (as in Mariah or Drew.)
    So to her, Aaron and Erin sound identical, as do Carrie and Carey.
    I keep saying it is a regional thing, becasue my family speaks the same way that I do...
    She just thinks we are wierd...

  17. I don't think lynneguist is right to say that the Italian /a/ in pasta corresponds to BrE PALM vowel (= father) rather than BrE TRAP vowel (= bat). BrE (including RP) TRAP vowel is no longer the [ae] vowel of the textbooks, or the [ae] vowel of AmE, but is very like Cardinal vowel [a], which is also the kind of short, low, not very front vowel found in Italian pasta, and in Spanish, French, Greek and many other languages --it's a very common vowel type. And the PALM vowel in BrE is long, while the vowel in Italian pasta is short. So the TRAP vowel is a pretty good match for 'international' /a/. For Americans, however, the TRAP vowel is too front and too raised (and in some of its contexts, too long) to be a good match for the vowel of It. pasta, whereas the unrounded low back POT vowel makes a better match.

    BrE non-rhotic speakers are inclined to say that the PALM vowel 'has an R in it' because the name of the letter R is exactly the PALM vowel.

  18. I take your point, Max, that none of these vowels exactly matches the Italian, but when our Italian colleague says pasta and latte it still sounds more like the American version to me than the British version does!

    But I'm not a phonetician...

  19. Maybe you should ask your Italian colleague which English pronunciation he thinks sounds closer.

    There may also be differences between North/South Italy and Iberian/American Spanish accents, which could influence which English vowel is closest to the Romance vowel.

  20. khakhi/car key - no difference in the 'ah'. focus on the 'i/ee' khakI and car kEY. In the latter, the 'ee' is more stretched.

  21. Mollymooly--good point. I've asked our Italian colleague, who has lived in both the US and the UK. She said the American pronunciation is 'definitely' closer to the Italian for her.

  22. I'm a native speaker of BE but also speak fluent Italian and to me the vowel in the AE pronunciation is much too long, the vowel in the BE is a bit too short. I have always thought the BE was closer to the Italian but I must defer to your Italian colleague if she has experience of both.

  23. An update on the Italian situation. I had lunch with my Italian colleague today and asked her the pasta question again (the first time I'd asked her by e-mail). After asking me to pronounce the AmE and BrE versions again (to the extent of my ability to pronounce the BrE version like a native) and she said "I can't hear much difference between them!"

    To me, there was a big difference. But if you come from a language that doesn't have an /a/-/ae/ distinction, they probably all flow together...

  24. Lynne,

    Wouldn't one of the ways to figure out if there is a "perceived" post-vocalic r is to determine whether the vowel in words like firm or fern or form are nasalized?

    I know in my Upper Midwest American they aren't because the "r" blocks nasalization.

    Do those accents that drop the post-vocalic "r" nasalize the vowel? Of course, this could be accounted for by a rule-ordering.

  25. Good point, Bob. But as I've said before, I'm no phonetician! Someone else will have to investigate that...

  26. Lynne,

    I am not a phonetician either, but you can do a simple test.

    Close off your nose with your thumb and index finger.

    Say feet. No pressure in the nose.

    Say fiend. Notice the pressure in the nose when articulating the vowel.

    Now, test your pronunciation of farm or fern. I don't get any pressure when articulating the vowel.

    If the Better Half has an r-less dialect, ask him if he has pressure when articulating the vowel in farm or fern or relevant words.

  27. BH is in America and I'm home alone, so someone else is going to have to do the test...

  28. Both my AmE wife and I (BrE) get a soft vibration using the close the nose test :P

    However that vibration comes with the consonant after the vowel. It happens wit farm, fern. fiend - but not with feet.

    Neither of us could feel any nasal differences against the vowel in any of the words.

    So as research - that tells you little more than the test isn't universal, which I suspect you knew.

  29. "Latte" has always annoyed me... I can say it (I think accurately) in an Italian accent, but that would sound pretty pretentious in Caffe Nero. The /a/ sounds to my (BrE) ear like a cross between the "trap" vowel and a short version of the "bath" one - perhaps "lartay" is slightly closer than "lattay". Maybe I should just order tea.

  30. The "a" in "latte" is almost identical to the northern English long "a", as in "far" (although not remotely like American or Southern English pronunciations of a long "a"), as is the "a" in "-are" endings etc. The "a" in "pasta" is the same as the northern short "a", as in "bat" (and in "bath" too, up here). My Italian teacher said the latter spot-on. They talk broad in Europe.

  31. This debate (ant vs aunt) always seems funny to me - as a Scot, these sounds are the same. There might be a miniscule difference in length, but it's a matter of cadence. I would never consider them different vowel sounds.

    Some English accents (the posher ones) do have an audible distinction, but it's nowhere near as exaggerated as in American English - if you do push it that much, as you point out, it sounds like a dropped 'r' ('parsta').

    As someone who grew up listening to English accents, I can hear the dropped R very clearly. Otherwise I would struggle to understand! To my ear, however, the dropped R is completely absent at the end of a word - "tuba" = "tuber", they're distinguished by context. This does vary with accent, mind you - in Scouse, for instance, there's an audible difference ("tubah" Vs "tubeh").

    Not all English accents re-insert the dropped R when following with a vowel. Some just insert a short glottal stop ("the tuba. is yellow")

    Sounds often fail to map between different UK dialects. Many dialects have their own unique sounds (for example, the Scots short and long "u") which have no equivalent in others. "English" is not really one language - it's a family of languages, mish mashed from all sorts of roots, and irregular in all sorts of fascinating ways.

    1. Scotland is different. Scottish accents have a vowel inventory unlike the rest of English. The only accents to share it are in neighbouring North Northumberland and the 'colonial outpost' as it were of Ulster.

      In the vast generality of English accents, there is a distinction between the vowel in words like TRAP and the vowel in the small group of word like PALM. For the much larger group of words like BATH there is a split between regional accents — and in some cases within a particular accent.

      Outside England, accents each assign one vowel value or another to subsets of BATH words with no consideration of what is 'posher' and what isn't.

      In England, there's a general difference between Northern speakers who pronounce BATH words with a TRAP vowel and Southern speakers who pronounce it with a PALM vowel. But an additional social factor compels some Northern speakers to see the BATH pronunciation as more prestigious, and modify their accent accordingly. To the speakers around them, this may be perceived as pretentious or posh.

      I, for example grew up in Nottingham (East Midlands of England). Children around me pronounced BATH words with a TRAP vowel. But I was in a middle-class family. My father had gone to a boarding school and my mother had paid for elocution lessons to lose her Swansea (South Wales) accent. Money well spent — she held down well-paid work despite having no qualifications in a time of high unemployment. So I pronounce aunt and ant with completely distinct vowels.

      The American complication is that something happened to the TRAP vowel in accents of New England, spreading to places like Detroit and Chicago. Before certain consonant sounds — including n — and after certain consonant sounds, the vowel may be longer and closer to the DRESS vowel. Or it may not — it differs from accent to accent.

      The difference in the pronunciation by one speaker between, say, the vowel of tan and the vowel of tang is more obvious to ears outside North America — hence your description of it as 'exaggerated'.

      The choice of inserted r in The tuba is yellow is not a matter of accent. Some speakers — myself included — always include a so-called intrusive R. Some never use the pronunciation. Many, I believe, pronounce the r without realising it. In casual speech, that is. In careful speech they use what they consider the more 'correct' pronunciation — the one that corresponds to the spelling. One choice or the other may be observed no matter what the accent. It's more controversial in the prestige RP accent because many RP speakers consider it an 'educated' accent.

      My accent is RP (more or less), so I make no distinction between aunt and aren't.

      There are powerful arguments for considering Scots a different language, but the prevailing tongue in Scotland is English with a Sottish accent. Yes, there are differences from other Englishes, but they relatively few and relatively superficial.


The book!

View by topic



AmE = American English
BrE = British English
OED = Oxford English Dictionary (online)