I am misunderstood when I say can't--almost as often as I'm misunderstood when I do my modern dance interpretation of Aspects of the Theory of Syntax. Many BrE speakers have difficulty hearing the difference between most AmE pronunciations of can and can't.

In both dialects, it's not really the 't' that helps us tell the difference between the positive and negative words--that sound is mostly swallowed at the end of the word. It's the vowel that makes the difference. In standard BrE, the vowel quality is the clincher. Can rhymes with pan but can't is pronounced like the AmE pronunciation of Kant (see the comments for more discussion). In other words, the vowel in can't is considerably further back in the mouth than the vowel in can. (One feels the need to mention the old chestnut: Kubla Khan, but Immanuel Kant--but note that Khan and can are pronounced differently.)

In AmE, the vowel quality is very similar between the two forms, but the length of the vowel in can't is shorter. (By a 'shorter' vowel, I literally mean 'shorter'--i.e. not pronounced for as long.) I explained that fact to my students yesterday, and once they knew that they got much better in the "which one am I saying?" test.

Pronouncing can't the other way is a favo(u)rite way for British singers to make themselves sound more American, and for American singers to make themselves sound more British. (I suppose either identiy has some cachet, depending on what kind of sound you're going for.) I've got to run to London to celebrate some fellow Librans (yeah for us!), so will leave it as your homework assignment to identify one example of each!


  1. "In standard BrE": I think you mean "in Southern BrE". A large chunk of England, and all of Scotland, rhymes can and can't with pan and pant - all the one vowel. Of course, in parts of Edinburgh, the vowel in question sounds more like "pent".

  2. I have never heard a person from the south-east of England pronounce 'Kant' and 'can't' the same at all. 'Kant' is like the first syllable of 'canter', and 'can't' rhymes with 'aren't' (in my opinion). They are quite similar in westcountry type accents, though the vowel is still not identical.

  3. ps - dearieme, yep, standard BrE does mean 'standard southern English English', pretty much! :)

    (it's not a value judgement though)

  4. I'm agreeing with those above me - I'd rhyme can with pan but can't with aren't. And it wouldn't rhyme with Kant, although that's maybe becaose of the teachers I had. I had three different pronunciations of Iraneus in a two year A level course!

  5. I think we are much more attuned to the glottal stop (the remnant of the 't' that is gobbled up by being in a coda) than we give ourselves credit for.
    In most registers of Australian English, a final 't' is very seldom released, but we tend to be responsive to the subtle sound of a glottal closure at the end of a syllable.

    I can't halp but think of Mick Jagger singing 'satisfaction'. He sings it with an AmE accent of course, yet the vowel length seems (to me) to be the same as it would be if the lyric were I can get no... (Of course the double negative serves to disambiguate, making it a less-than-optimal example, but ignore that).
    If the vowel length is indeed the same, then the only phonological clue as to the negative can only be the brief glottal closure between can and get.

    Grr. I tire of phonology. Any chance of some trans-Atlantic syntax issues, Lynne?

  6. I just want someone to explain the pronunciation of "pasta" (and also "Nicaragua").

  7. I should have said that southern BrE can't sounds like AmE Kant! (I'll change that now.) It does rhyme, for many BrE speakers with their pronunciation of aren't, as Ally says, but that doesn't help an AmE speaker know what it sounds like, since it doesn't rhyme with standard AmE aren't, which of course has an 'r' sound in it.

    So, dearieme (or anyone else who'd like to answer), would you say that the distinction between can and can't in the northern dialects is the same as that in American--i.e. is there a difference in vowel length?

    'Standard dialect' is regularly used in linguistics to mean 'the most unmarked dialect' (i.e. the one that's least considered to be 'a dialect' in layperson's terms) or 'the prestige dialect'. As such, it's the set of pronunciations that are usually taught to learners of the language (though it's becoming more common for AmE pronunciation to be taught in non-Anglophone countries--which is another contentious issue).

    As Ally notes, this is not meant to infer that there's anything linguistically 'better' about the 'standard', but to reflect some sociological relations among dialects. I hesitate to use the term RP, as that tends to be associated with some rather plummy pronunciations that are marked these days, though one can make a distinction between 'marked RP' and 'modern RP', as explained on this BBC wikipedia-like website. John Wells at University College London has a nice essay on RP (concluding that it should stay the target pronunciation for EFL learners) here.

    The BrE pronunciation of Kant takes us back to the discussion of how names should be pronounced. It wouldn't be pronounced like the first syllable in canter in German, would it?

    There are two problems with the hypothesis that we rely on the glottali{s/z}ed 't' at the end of can't, Aidhoss. First, word recognition rarely waits for the end of the word. This is shown through gating experiments, which demonstrate that the recognition point for a word is often quite a bit earlier than the end of the word. I've not seen a gating experiment using can't as a stimulus, but the information for distinguishing between can and can't is available to listeners well before the /t/. (And in a sentential context, other information will rule out it being other words like cantaloup(e) or Canterbury, etc.) Second, the /t/ at the end of can't will often be lost in context because of the following sound. For instance, if I say I can't tinker quickly, there's no release of the /t/ in can't before I go on to the /t/ in tinker. Much of the information about the stop (aka plosive) consonants that makes it to hearer's ears is in the vowels before/after them.

    Thanks to Aidhoss for doing the homework. (Another one: The Who's I Can't Explain.) But do we have any takers for Americans going for the southern BrE pronunciation?

    (Aidhoss, if you want a particular syntactic thingie covered, e-mail me! I've certainly been dealing with grammatical issues--i.e. verb complementation--lately.)

    Kathyf, pasta has to wait for another day, but it's definitely on my list!

    Apologies for the huge comment--a big catch-up after being off-line for *gasp* a whole 30 hours!

  8. Could someone please confirm my suspicion that in all dialects, the vowel sound in an unstressed "can" collapses to a schwa - for example, Kipling's "If you can keep your head when all about you". I don't believe would happen with "can't".

  9. I remembered as well that I used to work with a little girl who said 'I CAN'T!' as in 'canter', which to me said something interesting about child language acquisition. She knew the word 'can' and therefore it made sense to make the negative sound the same :)

  10. Rebecca, probably as likely an explanation for the girl's pronunciation is her exposure to that form either in a northern England/Scotland accent (you're up thataway, aren't you?) or an AmE one. Would have to know details of her exposure to the forms to conclude that she was pronouncing it that way on analogy with can. Remember--children are learning from the sounds of the words, not the looks of them, so it's very likely that they'd learn can/can't as separate forms to start out with (if they're pronounced with different vowels), not necessarily seeing them as two forms of the same word.

    Simon, you're probably right about all dialects being able to reduce the vowel in can to a schwa, so the question becomes whether it's also true that none can reduce it in can't. I don't know about that, but it seems likely that you're right because (a) the negated form would naturally take more sentential stress, thereby keeping a full vowel, and (b) pronouncing can't with a schwa sounds like a rude word--in which case I'm sure I would've noticed it happening!

    It wouldn't be foolproof for us to interpret any non-schwa-ful [kaen] as the positive version--since one can pronounce can with a full vowel (as one must in this sentence!). But I'm sure the reduceability of the /ae/ in can is one of the many factors that help us distinguish the words in listening.

  11. This reminds me of Haylee Mills in the original "Parent Trap", when Sharon has to teach Susan how to talk when she goes to Boston. "Aunt, shan't, can't." I never got this when I was little, because I pronounce 'Aunt' that way anyway, and lots of people of my parents' generation pronounce 'can't' that way too. And really, who uses shan't?!

  12. I'm too ignorant to answer your question directly, but I've just about got a point to make. "all of Scotland, rhymes can and can't with pan and pant - all the one vowel. Of course, in parts of Edinburgh, the vowel in question sounds more like "pent". Well, in truth I've not lived in all of Scotland, but in all that I know the vowel is the same in those four words, whether it's the usual Scots English "a" sound or the Morningside "e" sound (which my mother used!), or in the Southern Scots that I used in the playground as a boy, where it's roughly the "aw" sound, as in 'a crow caws'. In other words, in my experience almost all Scots use the same vowel in the four words, even if it's a different vowel in different parts of the country or at different ages. As for "Aunt", it's identical to "ant", at least to the level of scrutiny that my ears are capable of. And to my mother, both were close to "ent".

  13. Funnily, I say "Kant" like "can't" - both rhyme with "pant" or "ant" - or "aunt" for that matter (and I don't know why, since no other word with "aunt" in it sounds like that for me). And I know better with Kant but learned it before I did, so I have to consciously remember to say Kahnt. Also funnily, I pronounce Khan to rhyme with "con", so "Kubla Khan but Immauel Kant" doesn't work for me because of "Khan"...

    (ps - born in Tennessee to parents of mixed stock - Ohio and Alabama)

  14. ridger, I was having the same problem as you with the Khan/Kant distinction. I suppose it's a Midwestern United States thing, because I've grown up in Illinois.

    I've really enjoyed this post and discussion, if only because it's given me an excuse to say, "can, cahn, ken, kent, cant, cahnt" quietly to myself under my breath as coworkers walk by. I can't help but be reminded of the great exchange in the film, "Singing in the Rain" where Jean Hagen's character Lina Lamont is getting a diction lesson. When instructed to speak using "round tones," the following exchange occurs:

    Instructor: Now, let me hear you read your line.
    Lina Lamont: And I cayn't stand'im.
    I: And I can't [pronounced cahnt] stand him.
    LL: And I cayn't stand'im.
    I: Can't.
    LL: Cayn't.
    I: Caaaan't
    LL: Cayyyyn't

    A perfect example for this discussion!

  15. But Ridger, it's funny because the words are pronounced with the wrong vowels!

    Thanks for the bit o'dialog(ue), Anonymous.

  16. Really, really late on this one, but just wanted to point out that, at least to my untrained ears, the difference between AmE can and can't seems to lie in the length of the N sound, not the A. The N in can sounds a little more voiced than in can't.

    However, I do agree with whoever commented about the schwa being used in "can".

    PS. My husband and I got married the day you posted this entry. =)

  17. Really late here too. I find it interesting that Aidhoss brought attention to Mick Jagger"s use of the word satisfaction but neglected to note his uninquely southern AmE pronunciation of can't in the song. Here the vowel changes but instead of sounding like "Kant" it is lengthened as in "Cain". This form is common in rural portions of the southern US and is typical of the dialect of the bluesmen Jagger is imitating.


The book!

View by topic



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