Glottal stalking: Cockneys everywhere

« previous post | next post »

Today's SMBC starts with a little lesson in phonetic dialectology:


And then takes it in a somewhat unexpected direction:

Update — as several people have pointed out to me, the original strip said that a glottal stop "is made by halting air flow with your tongue". The mistake was corrected early in the morning of May 8, after Zach Wiener received "10,000 emails about how I'm wrong today", and "was killed by a horde of angry linguists" (not including me).



52 Comments

  1. Pflaumbaum said,

    May 8, 2012 @ 12:48 pm

    It's not just /t/ that is realised as a glottal stop in Cockney (or lots of other English and Scottish accents). It can replace /k/ in e.g. Buckingham and /p/ in e.g. supermarket, and that's before you get to all the pre-glottalised final consonants.

    Can a glottal stop replace /d/, rather than /t/, in certain American accents? I seem to hear it in AAVE, as in the much-memed "Oh no you di'int". If so, what are the environments in which it happens? Would you get it in, say, suddenly?

  2. Ran Ari-Gur said,

    May 8, 2012 @ 12:50 pm

    The comic implies that in Standard American English, glottal replacement of /t/ (i.e., production of /t/ as [ʔ]) only happens before a consonant, but I'm not sure that that implication is true. The English Wikipedia article on glottalization says (1) that pre-glottalization or glottal reinforcement of /t/ (i.e., production of /t/ in a way that incorporates both [ʔ] and [t]) happens commonly in syllable codas, and (2) that glottal reinforcement and glottal replacement are in partially free variation. That accords with my personal impression.

  3. Chandra said,

    May 8, 2012 @ 12:57 pm

    I don't think this will work very well as a trap. People who already know about glottal stops will already be aware of this and it will sound perfectly unremarkable, whereas people who don't know about glottal stops will insist that they really do hear a /t/ there. The same way that people who don't know what a schwa is will insist that they pronounce the last syllable differently in the words mother, hangar and realtor.

  4. Mike Aubrey said,

    May 8, 2012 @ 1:07 pm

    Confusion of glottal stops with English's word-final unreleased stops?

  5. Alexis said,

    May 8, 2012 @ 1:20 pm

    @Chandra,

    I don't know, when I first learned about glottal stops my professor used "Batman" as an example. And for at least a month I was mesmerized by the word. :P I could totally hear it! Anyway, not a big deal, but I think newbie linguistics students may get stuck on the glottal stops in common/cool words at least for a little while.

  6. chh said,

    May 8, 2012 @ 1:26 pm

    Mike-

    I was kind of wondering the same thing. My coronals undergo place assimilation to the following consonant, and final voiceless stops are unreleased, but I don't think it makes sense to call either glottal stops because there's still another gesture going on.

    However, there are definitely dialects of American English (at least on the East Coast where I live) where syllable final 't' really is a glottal stop, and "batman" would be pronounced exactly as the author of the comic suggests.

  7. Kenny said,

    May 8, 2012 @ 1:34 pm

    @Mike Aubrey
    I was wondering about that. A few weeks ago on NPR's A Way With Words, the male host said that the t's at the ends of words in American English were glottal stops. In my Into to Linguistics class we talked about those as unaspirated or unfinished allophones of the "normal" sound. We never mentioned glottal stops.

    Is there a difference? Or are there sometimes coarticulations of glottal stops and unfinished "letters"? I could probably be convinced that I feel the glottal stop in action on word-final t's and d's (whether it's happening or not) but not on word-final p's and b's (again whether that's true or not).

  8. Chandra said,

    May 8, 2012 @ 1:55 pm

    @Alexis – I wouldn't count newbie linguistics students in my cynical assessment above, as I would expect them by nature to be more interested in paying rapt attention to the phonemes people utter. :) (I certainly remember spending most of my first semester totally blanking on the content of people's conversations because I was so absorbed with the new-found task of mentally transcribing their speech into IPA.)

  9. richard howland-bolton said,

    May 8, 2012 @ 2:17 pm

    I'm so glad he corrected the original first panel. Early this morning* he had the glottal stop made with the tongue!
     
      
     
    * 4:00 am Central early!!

  10. Pflaumbaum said,

    May 8, 2012 @ 2:26 pm

    @ Mike Aubrey, Alexis –

    Is there no pre-glottalisation or glottal reinforcement before your unreleased unvoiced stops?

  11. Tim Martin said,

    May 8, 2012 @ 2:28 pm

    Can someone explain to me why these don't become unreleased /t/'s? How can you tell whether a word employs a glottal stop or an unreleased /t/ in SAE?

    Which sound is present in words like "button" or "pizza"?

  12. Tim Martin said,

    May 8, 2012 @ 2:29 pm

    Oh, I see a few people have asked a similar question already. Incidentally, I did email Dr. Liberman about this this morning, and he apparently wasn't inclined to answer the question via email or blog post. Hopefully someone will!

    [(myl) A lot depends on what kind of [t] you're talking about. There are two ways to cut off voicing rapidly in a final consonant: (1) spread the glottis, or (2) constrict the glottis. If a syllable-final [t] involves a glottal stricture coincident with the oral closure, call it a "glottalized [t]". This is a common pattern in American English for syllable-final voiceless stops before a syllable-initial consonant (though devoicing by glottal spreading can also happen). In neither case is the syllable-final stop in a …VCCV… sequence likely to be released, in fluent speech, independently of the release of the following syllable onset.

    Now, it's possible for the oral closure associated with the consonant to be weakened or elided completely. If the oral closure is completely gone, and the final consonant is glottalized, then what's left is a glottal stop. (As long as the constriction for the following consonant doesn't occur early enough to take over…) As the cartoon suggests, the transformation of a glottalized syllable-final [t] into a glottal stop is a pretty common situation in fluent American speech.

    At this level of description, it's not very helpful to talk about segment sequences. It's more useful to think of an articulatory score, in which different articulators are changing state over time in a way that's independent but coordinated, like the performances of different instruments playing together in an ensemble.]

  13. diogenes said,

    May 8, 2012 @ 2:38 pm

    Pflaunbaum – the suppression of /d/ was one of Kingsley Amis's bugbears – many of his later novels have characters who are transcribed as saying "vogka" rather than "vodka".

  14. Tom Recht said,

    May 8, 2012 @ 2:41 pm

    I think there's a change underway in the pronunciation of words like 'button', 'Britain', etc., at least here in California. I'm used to hearing and pronouncing these with unreleased [t̚] or glottal stop followed by a syllabic nasal, but lately I've been hearing college-age speakers flap the /t/ instead, just as in 'butter' or 'buttock'. It's a natural expansion of the allophony rule. I wonder what its distribution is – has anyone else heard this?

  15. Jimbino said,

    May 8, 2012 @ 3:07 pm

    It appears that metal, mettle, medal and meddle are homonyms in American.

  16. Howard Oakley said,

    May 8, 2012 @ 3:27 pm

    Bring on the Caucasian languages with their rich range of glottalized consonants, as in Standard Georgian, perhaps?
    Georgian has some notorious tongue-twisters that would make even a Cockney stumble.

  17. David Eddyshaw said,

    May 8, 2012 @ 3:32 pm

    @diogenes:

    Yes indeed; in Amis' "The Green Man" there's a trendy Anglican vicar who gets this treatment applied pretty consistently, along the lines of "sip back" for "sit back" etc etc.

    I remember being irritated by this when I first read the novel in my teens, as a bit of introspection and experiment revealed that Amis was just doing a more sophisticated version of the bad habit of writing "sez" and "duz" for "says" and "does", not to indicate anything special about pronunciation but to express an authorial attitude about the speaker. I'm pretty sure Amis disapproved of trendy Anglican vicars (who doesn't, poor things?) but even now I'm not clear how he felt that their bad attitudes spilt over into linguistic depravity exactly.

  18. Sevly said,

    May 8, 2012 @ 4:09 pm

    Tom Recht said,
    I think there's a change underway in the pronunciation of words like 'button', 'Britain', etc., at least here in California. I'm used to hearing and pronouncing these with unreleased [t̚] or glottal stop followed by a syllabic nasal, but lately I've been hearing college-age speakers flap the /t/ instead, just as in 'butter' or 'buttock'. It's a natural expansion of the allophony rule. I wonder what its distribution is – has anyone else heard this?

    I'm an example of it. I remember in my intro linguistics class where my teacher started telling us that the Standard American pronunciation of kitten was [kɪʔn̩] and of button, [bʌʔn̩] and I was just stunned. I have always pronounced those words with a flap, as kitten [kɪɾn̩], button [bʌɾn̩], but interestingly enough I vary when it comes to Britain [bɹɪɾn ~ bɹɪʔn̩].

  19. dw said,

    May 8, 2012 @ 4:40 pm

    The place where I find glottal stops most striking in North American accents is between a nasal and a syllabic nasal, as in "mountain" or "Clinton".

  20. Bloix said,

    May 8, 2012 @ 5:24 pm

    Sevly, for you kitten and kiddin' are homonyms? Really?

    My mother, an elderly New Yoker, says bah'l for "bottle." We make fun of her for it.

    In "standard British" (estuary English?), "Batman" is not merely "Ba'man," it's "Ba'mam," which I find truly weird.

  21. Jim said,

    May 8, 2012 @ 5:31 pm

    "It appears that metal, mettle, medal and meddle are homonyms in American."

    They aren't for me or anyone else around me – West Coast, Bay Area and then Seattle – but they are very close. Voicing is the difference, but it lasts for a very short time. Probably most speakers of the European varieties of English mentioned in the first comment would have a very difficult time distinguishing them.

    Sevly,
    "I'm an example of it. I remember in my intro linguistics class where my teacher started telling us that the Standard American pronunciation of kitten was [kɪʔn̩] and of button, [bʌʔn̩] and I was just stunned."

    I agree with you. Where was this teacher from in the States?

    " I have always pronounced those words with a flap, as kitten [kɪɾn̩], button [bʌɾn̩], but interestingly enough I vary when it comes to Britain [bɹɪɾn ~ bɹɪʔn̩].

    Not me. Because those were historically 'n' they keep the 't' a stop, unreleased, whereas "bottom" or "getting" I pronounce with a flap

  22. Chandra said,

    May 8, 2012 @ 5:43 pm

    Re. metal, mettle, medal, meddle – They're homonyms for me (Canadian) unless I'm enunciating carefully.

  23. John Lawler said,

    May 8, 2012 @ 6:07 pm

    According to the Official Rules, as taught in at least some linguistics classes, English stressed vowels before voiced consonants are longer than those before voiceless ones, even if the voicing disappears due to neutralization.

    And /d/ and /t/ both neutralize to [ɾ] following a stressed vowel and preceding an unstressed one.

    This would make metal and mettle homophonous, as ['mɛɾḷ], while meddle and medal would also be homophonous, as ['mɛ:ɾḷ], but would be distinguished by its slightly longer stressed vowel from metal and mettle.

    No doubt this is true some of the time, for some English speakers.

  24. Pflaumbaum said,

    May 8, 2012 @ 7:30 pm

    @ Bloix

    "In "standard British" (estuary English?), "Batman" is not merely "Ba'man," it's "Ba'mam," which I find truly weird.

    Maybe what you mean is RP? There's no 'Standard British', but RP is the prestige accent of England. Estuary is not RP though. And I'm not sure that either of them have [m] at the end of Batman, except when the following word begins with a labial, as in Batman Begins (which might be something like ['bæʔtmæmbɪg'ɪnz] in RP and ['baʔmambɨg'ɪnz] in Estuary).

  25. Tim Martin said,

    May 8, 2012 @ 9:10 pm

    @Dr. Liberman: Thank you! I think I understand. In the case of an unreleased /t/, the tongue resting against the alveolar ridge and cutting off the air stream is what you're calling the "oral closure," is that right? So that closure may or may not occur, but the glottis is also doing part of the work to cut off the sound from the previous syllable? This was not something I'd heard before.

    I think there's a change underway in the pronunciation of words like 'button', 'Britain', etc., at least here in California. I'm used to hearing and pronouncing these with unreleased [t̚] or glottal stop followed by a syllabic nasal, but lately I've been hearing college-age speakers flap the /t/ instead, just as in 'butter' or 'buttock'.

    This is something you'll also hear in the New York area. I've got a friend from Manhattan and a friend from Connecticut who both flap the /t/ before syllabic /n/. So they pronounce my name like "Mardin" (this has actually confused speakers who don't have this as part of their dialect, on at least one occasion).

  26. Tim Martin said,

    May 8, 2012 @ 9:13 pm

    I meant to add that these friends are around age 30.

  27. Hana said,

    May 8, 2012 @ 9:57 pm

    And to add to Tim's findings, i once caught an episode of "The Apprentice" and was so struck by the way that Ivanka Trump said 'Manhattan' that i actually remember it. It was just like that- 'Manhadden' – first time i'd encountered that pronunciation.

  28. michael farris said,

    May 9, 2012 @ 3:02 am

    John Lawler: "English stressed vowels before voiced consonants are longer than those before voiceless ones, even if the voicing disappears due to neutralization"

    Not true in my version of SAE. I do have a difference in vowel length before word final consonants with the vowel of bad being longer than that of bat, lab longer than lap and hag longer than hack.

    But the neutralization of t and d neutralizes this too and metal, mettle, medal and meddle are all homonyms in my idiolect.

  29. TonyK said,

    May 9, 2012 @ 5:00 am

    @ Bloix, Pflaumbaum: I agree with Bloix — Batman can be pronounced "Ba'mam", even at the end of an utterance. Especially at the end of an utterance, in fact — the lips are closed prematurely, before the word is complete.
    As for Buckingham, for some speakers it can (but needn't) sound the same as "butting them".

  30. John said,

    May 9, 2012 @ 6:37 am

    Bloix,

    Don't forget the Bel Paese, I'ly.

  31. Mr Punch said,

    May 9, 2012 @ 10:01 am

    "Metal" and "medal" are different for me.

    @Pflaumbaum – There are American accents in which glottal stops replace /d/. Medford, Massachusetts is notoriously called "Me' fah" by its inhabitants; visiting there recently, I was very nearly told to "have a goo' 'ay."

  32. Sevly said,

    May 9, 2012 @ 1:50 pm

    Bloix said,
    Sevly, for you kitten and kiddin' are homonyms? Really?

    Yes, sir.

    Jim said,
    I agree with you. Where was this teacher from in the States?

    Arizona.

  33. Pflaumbaum said,

    May 9, 2012 @ 3:11 pm

    @ TonyK

    I've never heard of this final [m] or picked it up in my own or others' speech. Is there any literature on it you could point me to?

    @ Mr. Punch

    Surely they don't come close to realising an initial /d/ as a glottal stop?

  34. Jimbino said,

    May 9, 2012 @ 4:24 pm

    It also appears that carat, karat, carrot and caret are homonyms.

  35. Jimbino said,

    May 9, 2012 @ 4:29 pm

    And what about air, heir, e'er, ere, hair & hare (cockney), and are (2.5 acre)?

  36. TonyK said,

    May 9, 2012 @ 4:38 pm

    @Pflaumbaum

    I'm still searching the web for corroboration of sentence-final lip closure converting [n] to [m]. But in the meantime, this 1946 article from the journal "American Speech" discusses the "lip-closure of finality" that I was trying to convey. If it can change "yes" to "yep", then surely it can change "ba'man" to "ba'mam".

  37. Bloix said,

    May 9, 2012 @ 5:36 pm

    Plaufenbaum:

    See,e.g., Kingsley Amis, Lucky Jim:

    "'I don't happen to be that type, you sam.'
    'I'm not Sam, you fool," Dixon shrieked. This was the worst taunt of all."

  38. Pflaumbaum said,

    May 9, 2012 @ 6:56 pm

    @ Tony K –

    It's easy to see why yes and no can acquire a marker of 'finality', but less clear to see why Batman might.

    Yep and nope are also of course used in dialects well beyond England. They are quite pecuiar in that, as discussed on here before, they seem to virtually always appear as one-word utterances ; and also, in my pronunciation anyway, the /p/ seems to be ejective, an allophone (if it is one) that appears nowhere else in my speech.

    @ Bloix –

    That's Bertrand's own particular vocal mannerism, and also nothing to do with final /n/ – 'sam' represents 'see'. Realising /i/ as an open/nasal vowel might have been an affectation Amis heard – or thought he heard – among social climbers in the early fifties, but I'm pretty sure it's not general to RP or Estuary today.

  39. Bloix said,

    May 9, 2012 @ 8:06 pm

    Plaufenbaum –
    Bertrand drawls his ee into a lazily "aristocratic" eh or ah sound (sorry, I'm a rank amateur and don't know IPA) and then arrogantly snaps his lips shut – "you sah (snap)." Amis's representation of it as "sam" is the only genuinely hilarious example of dialect spelling that I'm aware of.

    I agree that it's different from the move that produces Ba'mam and many other final M sounds. I listen to the BBC World Service from time to time (it's on at night on most PBS stations here) and the m's drive me nuts. For example, the word "in" routinely becomes "im." E.g., "Im Brussels today … " What's going on, it seems to my untrained ear, is that the speaker can't be bothered to form the n and shapes his or her mouth to the initial letter of the following word, which produces an "m" sound.

  40. Pflaumbaum said,

    May 9, 2012 @ 8:40 pm

    Oh yes, the /n/ in in Brussels is definitely realised as [m], because of the following labial. That's what I meant when I was saying it would happen in Batman Begins. The classic example is 'Londom Bridge'.

    But this isn't particular to English accents. It may jump out at you when you're listening to the BBC because it's a foreign accent, but I'm pretty sure that Americans do the same in fluent speech. /n/ will also normally assimilate to [ŋ] (the sound at the end of long before /k/ and /g/, so 'John Coltrane' will generally be pronounced 'Jongcoltrane'.

    Well, my half-arsed attempt at substantiation is a wash: Obama here seems to say American President with an [n] at the end of American, but immediately afterwards American people with an [m].

  41. Mark F. said,

    May 9, 2012 @ 10:10 pm

    Earlier this evening I had a store clerk at Trader Joe's tell me they had swee'ened and unswee'ened apple sauce. I'd heard that accent before on the Planet Money podcast, from a NY/NJ area person.

  42. Bloix said,

    May 10, 2012 @ 12:34 pm

    I say swee'ened. Also Mar'in and ki'en. And sometimes I say som'in. That one drives my wife up the wall.

  43. Lewis said,

    May 10, 2012 @ 9:54 pm

    (Disclaimer: I'm not actually a linguist. I'm only a very enthusiastic amateur.)

    Like Mike I think people may be getting glottal stops confused with unreleased (or nasally released) plosives. I speak a variety of "General American" and for me the "t" in "Clinton" and "mountain" is not a glottal stop, but instead it is a "t" with nasal release. The "t" in "cat" is often an unreleased voiceless alveolar plosive for me. But in both cases, the tip of my tongue touches the roof of my mouth (at the alveolar ridge) during the "t". If they were glottal stops, my tongue would not touch the roof of my mouth, right?

    To my ear, glottal replacement is not nearly as common in America as many people seem to think. I think this is a misconception, which has for some reason become fairly widespread in recent years. I have heard glottal replacement in some northeastern (especially New York City area) speech and some AAVE, but that's about it. I don't think so-called "glottal reinforcement" is that common either. It strikes me as very English, if it sounds how I imagine it to sound.

  44. Adam said,

    May 11, 2012 @ 2:57 am

    Slightly related: today's Dilbert cartoon has a rhyme based on American flapping.

  45. dw said,

    May 11, 2012 @ 9:25 am

    @Lewis:

    I speak a variety of "General American" and for me the "t" in "Clinton" and "mountain" is not a glottal stop, but instead it is a "t" with nasal release.

    That's a possible pronunciation, but not one I hear very often in person (in Northern California) or in the US media. Interestingly, some British accents (such as RP, or indeed my own accent) have alveolar "t" with nasal release in words like "button", but _not_ where there is a /n/ before the /t/, as in "Clinton".

    But in both cases, the tip of my tongue touches the roof of my mouth (at the alveolar ridge) during the "t". If they were glottal stops, my tongue would not touch the roof of my mouth, right?

    Not necessarily. What matters is what part of your anatomy is actually stopping the airflow from your lungs. If the pressure from the lungs is actually stopped at the glottis, then it doesn't matter what your tongue is doing: you're performing a glottal stop.

  46. Glenn Bingham said,

    May 11, 2012 @ 12:12 pm

    @Pflaumbaum

    In African American speech, the rule for replacing phonemic /t/ with a glottal stop is generally extended to include /d/ as well. (I haven't worked through all the implications of this, but it seems to hold up so far.) So the difference between 'bet' and 'bed' is only the length of the vowel. Both end with a glottal stop.

  47. Glenn Bingham said,

    May 11, 2012 @ 12:22 pm

    @Kenny

    In our Mid-Atlantic US English, the tip of tongue does not venture near the alveolar ridge when pronouncing /t/ in such words as 'bat' or 'eat.' It comes out a glottal stop. Indeed, in my speech, the tongue lies near the base of the lower teeth (or what's left of them).

  48. Pflaumbaum said,

    May 12, 2012 @ 11:25 am

    @ Glenn Bingham –

    Thanks. But what is the rule in AAVE? As well as in final position as in your examples, and before syllabic /n/ and /l/, would some speakers have a glottal stop in, say, bitter? Surely not in bidder?

  49. Adam said,

    May 12, 2012 @ 3:04 pm

    My old man's a ba'man
    He wears a ba'man's hat…
    (sorry, I'll get me coat)

  50. Jack said,

    May 12, 2012 @ 6:00 pm

    Not a linguist, not even a qualified amateur, but as a lifelong Nu 'nglennduh, I can confirm that it's Me'fuh, Mass, just like it's Be'fuh, Nu 'Ampshuh, or even Paw'smuhth, Nu 'Ampshuh.

    Also: ki'in, mi'in, mow'in, baw'l, Bill Cli'in, swee'ind, carrih', parrih', but not meh'l, which is medh'l, but which is different from medul

    for: kitten, mitten, mountain, bottle, Bill Clinton, sweetened, carrot, parrot…and metal, which is different from medal.

  51. Julie said,

    May 14, 2012 @ 1:05 am

    I don't know the rules of AAVE, but I hear "Broa'way" fairly often in my neighborhood.

    For me, kitten and the like are usually pronounced with a glottalized t and no vowel. Words ending with t are another story. At least sometimes, it's a glottal stop. Not sure in what contexts exactly. I know I often pronounce my husband's name (Walt) with a glottal stop, but the dark l preceding may make a difference there. I'm a native of Northern California.

  52. BenHemmens said,

    May 20, 2012 @ 3:21 am

    A' the'en' o' th' dee, i's aa' abou' pu'in thi ba' i' th' bac' o' thi ne'.

    (The point of football, as often recited in Scotland. Very apt for last night's Champion's League final).

RSS feed for comments on this post