Language and identity

« previous post | next post »

Rebecca Tan, "Accent Adaptation (On sincerity, spontaneity, and the distance between Singlish and English", The Pennsylvania Gazette 2/18/2016:

The most difficult thing about speaking in a foreign country isn’t adopting a new currency of speech, but using it as though it’s your own—not just memorizing your lines, but taking center stage and looking your audience in the eye. It is one thing to pronounce can’t so that it rhymes with ant instead of aunt, but a whole other order to do that without feeling like a fraud. […]

Lately I’ve been wondering if I’ve taken this whole language situation a tad too personally. Till now, I have kept my Singaporean inflection close at hand, for fear that attempts at Americanisms will be wrong—or, worse, permanent. Yet I am beginning to feel myself grow tired of this stage fright, tired of this senseless preoccupation with the packaging of ideas rather than the ideas themselves. Away from all these theatrics, the simple facts are that I am 9,500 miles away from home, and will be for four more years. I came here looking for change, and the words forming in my mouth to accommodate that change are not jokes, lies, or betrayals. They are real, not strange, and they are mine.

Read the whole thing.


  1. Guy said,

    April 16, 2016 @ 4:30 pm

    On the topic of can/can't, I often can't tell tell which word non-native speakers are using, and part of the reason, I think, is misleading spelling together with insufficiently detailed education on the way phonemes are realized in native accents. In my accent, which I think is pretty close to what might be considered "standard" American, can is usually something like [kə̃n], and can't is usually something like [kẽə̯̃ʔ]. The /n/ phoneme is realized phonetically by raising and nasalizing the phonetic representation of the /æ/ phoneme, and the expected [n] portion is essentially clipped off by the early glottal stop preparing for the /t/ phoneme, and the /æ/ phoneme isn't even present in unemphasized can; the vowel is /ə/. Non-native speakers often only distinguish these two crucially different words with a difficult-to-discern [t] segment that doesn't affect the realization of the earlier phonemes in the way that is usual for the English /t/ phoneme.

    Part of the difficulty in non-native speakers learning this is that most native speakers aren't even consciously aware that these words are distinguished in the way that I just described, so it's hardly fair to expect non-native speakers to easily learn it.

  2. Pflaumbaum said,

    April 17, 2016 @ 12:19 am

    @ Guy-

    As a BrE speaker who spends a lot of time in the States, this is the thing I find by far the hardest to understand. (Or at least, it's by far the most salient thing, since the distinction is generally pretty crucial.)

    'Can' and 'can't' essentially sound identical to my ear in fast speech, especially on the West coast. Your analysis is interesting, I will try to look out for that.

  3. cameron said,

    April 17, 2016 @ 1:58 am

    The contrast between RP and US pronunciations of "class" is clear to me. But neither of those vowels matches the vowel in "gloss".

    The RP vowel in "gloss" is not used in American English. Do Singaporeans really pronounce "gloss" with the same vowel as RP "class"?

    I have actually been to Singapore. A few weeks in '04. Didn't notice this pronunciation oddity.

  4. wren romano said,

    April 17, 2016 @ 3:10 am


    I find those pronunciations quite far from what I'd call "standard AmE". In my own speech (which I lay no claims to being standard), I pronounce the [æ] as such in both words, and don't have any off-glide or lengthening in the negative. Your off-glide strikes me as a Southern trait, and I don't think I've ever encountered anyone neutralizing the vowel to schwa.

    I do agree, however, that in many dialects the "n't" cluster turns into a strong nasalization of the vowel followed by a glottal stop. Dunno if I'd go so far as to call it "standard" though. Without any basis, I feel like [kænt], [kænt̚], [kænʔ], and [kæ̃ʔ] are in free alternation depending mainly on clarity/rapidity of speech (along with differences of register which confound mere clarity/rapidity).

  5. Peter said,

    April 17, 2016 @ 4:25 am

    @cameron: I don’t know Singaporean English much, but to most Americans with the father–bother/lot–palm merger, the vowel of RP (and Singaporean?) class (i.e. the palm/father vowel) is the same phonemic vowel as their gloss (the lot/bother vowel), even though to RP (and Singaporean?) speakers, gloss and class are quite different. So I imagine that’s why Tan — writing for a UPenn newspaper, and presumably with general NE-American as her main point of reference for non-Singaporean English — described her Singaporean pronunciation of class as like gloss.

  6. Sky Onosson said,

    April 17, 2016 @ 8:40 am

    As far as this Canadian is concerned, can and can't are often so similar in rapid speech as to be completely indistinguishable; the only way to know which was said is to rely on context or to ask "Did you say can or can't?"

  7. Karen McM said,

    April 17, 2016 @ 10:21 am

    @wren romano: I neutralize the vowel to a schwa in "can" but not "can't" if I emphasize whatever comes before it, as in "I can think of an example of neutralizing the vowel."

  8. Kiwanda said,

    April 17, 2016 @ 12:22 pm

    "can't", "ant", and "aunt" all rhyme for me, just as "Mary", "merry", and "marry" do. I'm a bit surprised that doesn't seem to be true in Pennsylvania.

  9. Ellen K. said,

    April 17, 2016 @ 2:03 pm

    I didn't completely follow what Guy said in his comment, but one thing I did pick up on that I would say is true of my (relatively standard) variety of American English, is that the word "can" is ordinarily pronounce with a schwa (unstressed vowel), only being pronounced with /æ/ when stressed. So putting /æ/ in can where the word is not stressed, makes it sound like "can't". (When stressed, we don't use "can't" but "can not" or "cannot", so the positive and negative having the same vowel doesn't cause issues in understanding.)

  10. Ellen K. said,

    April 17, 2016 @ 2:12 pm

    @Kiwanda, though the /ɑ(ː)nt/ pronunciation is used some in the U.S., I suspect the author is thinking, not of the Pennsylvania pronunciation, but of the Singapore pronunciation, which presumably matches the British.

  11. monscampus said,

    April 17, 2016 @ 6:06 pm

    This film gave me a good idea of what kind of English is spoken in Singapore. I'd have no problem making myself understood there with my take on British English.

  12. monscampus said,

    April 17, 2016 @ 6:12 pm

    Sorry, it was this link to the full film I meant to post

  13. Guy said,

    April 18, 2016 @ 12:09 am

    @wren romano

    I'm a little surprised to hear you say that the vowel of can is not reduced to a schwa in ordinary usage. It's true that the vowel would always be /æ/ in citation form, that is, if you asked someone to read the word off of a card, but almost never in fluid speech without emphasis, just like the indefinite article is almost never /eɪ̯/ in ordinary speech.

    A raised or diphthongized /æ/ in all contexts I would say is a Southern trait, but not before /n/. Before /n/, /æ/ is typically raised in North American accents. It's easy to miss this if you have an Anerican accent because your brain is interpreting it in its phonetic context. It's the same reason you don't consciously notice that the vowel is nasal. I had to practice with a spectrogram before I was first able to fully convince myself that I raise /æ/ before /n/. Wikipedia isn't an infallible source, but it's immediately available, and you'll see that in the section under /æ/ tensing here that it's listed in the chart as raised before /n/ in most North American accents, including the one that it calls General American. This comports with my general observations.

  14. Guy said,

    April 18, 2016 @ 12:12 am

    I'm a lifelong resident of the San Francisco Bay Area, to the extent anyone is interested in geographic specifics.

  15. Vanya said,

    April 18, 2016 @ 7:49 am

    Sensible North Americans, mainly working class or rural New Englanders, do still distinguish "can" and "can't" by pronouncing "can't" to rhyme with "taunt" and "font". It is funny that in most of America that would be considered a pretentious upper class affectation.

  16. Kiwanda said,

    April 18, 2016 @ 11:15 am

    "@Kiwanda, though the /ɑ(ː)nt/ pronunciation is used some in the U.S., I suspect the author is thinking, not of the Pennsylvania pronunciation, but of the Singapore pronunciation, which presumably matches the British."

    Her note implies that "ant" and "aunt" don't rhyme in Pennsylvania, which surprises me a bit. (Otherwise, she'd talk about all three rhyming, not switching "can't" from one to another to go from Singapore to American pronunciation.)

  17. BZ said,

    April 18, 2016 @ 12:47 pm

    Although I'm not (completely) a native speaker of AmE (despite being here for 25 years), I can attest to others often asking each other in conversation about whether they meant "can" or "can't", so it's not just "sounds the same to me". This is in New Jersey by the way. As for "aunt", I've heard both pronunciations here, though the one that sounds like "ant" is somewhat more common, while the other one is often used to resolve ambiguities.

  18. Ellen K. said,

    April 18, 2016 @ 1:01 pm


    I can't see that it implies that. It merely indicates that they aren't pronounced the same for her. Since she isn't from Pennsylvania, and isn't writing about that word, it really doesn't imply anything about it's pronunciation there.

  19. andyb said,

    April 18, 2016 @ 1:13 pm

    I think the biggest distinction is that, for most Americans, the vowel in "can" (only the modal, not the noun) is /e/ or /ɛ/, not /æ/. And this means it doesn't undergo a-tensing, while the vowel in "can't" does. Secondarily (contrary to wren romano), the vowel in "can" is often reduced, but the vowel in "can't" rarely is.

    In California, I think the most common form of nasal a-tensing is to raise and lengthen, so "can't" vs. "can" becomes something like /keənʔ/ vs. /ken/. Some people definitely raise less but add an off-glide, as Gus describes, so it's /kɛjnʔ/ vs. /ken/. And California AAVE sometimes has the Southern glide-unglide thing, and may drop the t entirely instead of glottalizing it, so it may be something like /kejən/ vs. /ken/. In all those cases—plus most other American dialects—the easiest way to distinguish is the diphthong in "can't".

    For more evidence that "can" is /ken/ in American English:

    I think Gus is wrong that the citation form is /æ/: you can suppress a-tensing when carefully reading off a card, so "can't" may be /kænt/. But if you pronounced "can" as /kæn/, I think everyone would be sure you meant the noun, not the modal verb.

    According to what Vanya says, it sounds like "can" has a raised (and fronted?) vowel even in dialects with no nasal a-tensing.

    Even when the vowel is reduced, it's still raised—people who use the /ɪ/ or /ɨ/ vowel in plurals like "noses" also use it in unstressed "can"; syllables with /æ/ almost always alternate with /ə/ or lower.

    Try reading "I can see Ken", "I can see the can", "I can't see Ken", "I can't see the can", and imagining them in different accents.

  20. Sky Onosson said,

    April 18, 2016 @ 1:21 pm

    @andyb: the vowel in "can" is often reduced, but the vowel in "can't" rarely is

    I have to disagree with the latter part of that statement. In careful speech they are certainly distinguishable, but in many cases they are both reduced to the point where I'd simply transcribe them as [k] followed by syllabic [n].

  21. Guy said,

    April 18, 2016 @ 1:31 pm


    I would represent can phonemically in ordinary speech as /kən/, but since transcription standards vary, I should point out that I think /kn̩/ would be about as reasonable. I think can is usually homophonous with the second syllable of spoken in ordinary speech, but that wouldn't be a citation form because that pronunciation is only available for unstressed syllables. Some people might have /kɛn/ rather than /kæn/ for emphasized can (Citation form, polar emphasis, and ellipsis of the complement) I imagine that there's some variation here just as there is for than. I searched on Obama's "yes we can" slogan and I feel like /æ/ is what we usually hear, though I suppose I would need control samples from the same speakers to be sure. Because /æ/ is usually raised before /n/, the contrast between /æ/ and /ɛ/ in that position is a little subtle.

  22. Phil Ramsden said,

    April 18, 2016 @ 8:13 pm

    What beautiful, lucid writing, from such a young author.

  23. andyb said,

    April 18, 2016 @ 11:37 pm

    @Guy: "I would represent can phonemically in ordinary speech as /kən/, but since transcription standards vary, I should point out that I think /kn̩/ would be about as reasonable".

    In accents where the plural vowel in "noses" is realized as something like a slightly backed /i/, the vowel in "can" is as well. Think of the memes with the dialect-spelling "you kin do it" under a picture of Rob Schneider from every Adam Sandler movie. Of course there are plenty of accents where the plural is just a syllabic /z/, and in those accents "can" is also syllabic /n/ (or, if you prefer, /ən/). And some speakers alternate between the two (depending on something in the context? I'm not sure). Which makes me think it should be phonemically transcribed the same way as the plural vowel: /kɨn/, not /kn/.

    "Because /æ/ is usually raised before /n/, the contrast between /æ/ and /ɛ/ in that position is a little subtle."

    Sure, but a-tensing doesn't _just_ raise /æ/ before /n/ in most American dialects, it has other effects as well, and those other effects are how you can distinguish the vowel here.

    In most American accents, a-tensing replaces /æ/ before a nasal it with a diphthong (albeit different diphthongs for different accents). And, as you described phonetically for your own San Francisco accent, there is in fact a diphthong (an off-glide) in your "can't", but not in your "can". Which is exactly what you'd expect from /kænt/ vs. /ken/ in a California accent.

    It's _possible_ there's some other process going on here that makes "can" different, but I think it's just the simple fact that "can" is /ken/, not /kæn/, and then everything else is described by well-known general features of most American accents.

    And would you predict that carefully-pronounced "I can see the can" would have the same vowel for both "a"s?

  24. Kiwanda said,

    April 19, 2016 @ 12:14 am

    @Ellen K.

    From the UPenn gazette, by a first-year at UPenn: "It is one thing to pronounce can’t so that it rhymes with ant instead of aunt, but a whole other order to do that without feeling like a fraud."

    This implies that she wants to have the local accent, in Pennsylvania where she's going to school, and where "can't" rhymes with "ant" and not with "aunt", implying that "ant" and "aunt" don't rhyme. Unless there's some third place, neither Singapore nor Pennsylvania, that she's alluding to without explanation.

  25. Sky Onosson said,

    April 19, 2016 @ 8:14 am


    The way I interpreted that comment was that she was confronted with changing her pronunciation of "can't" but not her pronunciation of "aunt", which might be just because "can't" is a much more common word, and not because "aunt" isn't also pronounced with the same vowel in local speech.

  26. Kiwanda said,

    April 19, 2016 @ 4:43 pm

    ohhhhhhh. {\small never mind}.

RSS feed for comments on this post