Dinosaur intonation
« previous post | next post »
Today's Dinosaur Comic:
Mouseover title: "did i sit down at a keyboard to transcribe the 'i don't know' hum? well sir, i'll let this little three-note ditty here give a coy answer to THAT question"
I seem to have a few versions of this melody, mostly somewhat different from Ryan's. But I'm curious to see what versions are really Out There, in the anglophone world and elsewhere. (And in particular, I wonder if the melody is really "universal".)
So send please me recordings of your performances, in whatever language(s), with words or without. I'll report the results after a reasonable number of recordings have arrived.
Update — See the next installment, with a sample of Ryan North's own performances, at "Hummed 'I don't know'", 8/29/2021. And "More 'I don't know'", 8/31/2021.
Michael M said,
August 28, 2021 @ 1:35 pm
I'm pretty sure I have the same melody as Ryan, but then I'm from Ontario as well so that's to be expected.
Polish, I'm pretty sure, does not have an equivalent, and I would be very interested to here if any other language does.
[(myl) Please send a couple of recordings!]
Philip Taylor said,
August 28, 2021 @ 2:04 pm
Somewhat confused (a non-unusual state of affairs) — are there really idiolects (or topolects, or whatever) in which one can answer "I don't know" to a question by simply humming three notes ? If so, it would seem that I have never encountered (or perhaps recognised) someone who answers in that manner.
Y said,
August 28, 2021 @ 2:25 pm
Do you want a "neutral" tone as well? In other words, a bit of voice with what the speaker perceives as neutral/relaxed vocal cords?
[(myl) I don't think that would be especially helpful, but who knows?]
Michael Watts said,
August 28, 2021 @ 2:55 pm
It's part of standard American English. It's better thought of as one pitch contour than three notes; the contour is identical to that which accompanies the fully realized phrase "I don't know". All of the consonants have been omitted.
I think of it as part of a set with uh-huh and uh-uh – they are the three possible answers to yes/no questions, and all of these special forms allow you to answer without opening your mouth. (Opening the mouth is possible, but not required.)
[(myl) See "mmhmm etc.", 8/18/2018; and "Huh", 2/4/2010. Also Mark Liberman and Ivan Sag, "Prosodic Form and Discourse Function" (1974), which is directly relevant to the original post and the comic it's based on.]
Michael Watts said,
August 28, 2021 @ 2:57 pm
The "I don't know" contour is interesting to me in that it establishes English as separately exhibiting both lexical tone (the contour) and grammatical tone (the rising intonation represented by the "?" symbol).
Philip Taylor said,
August 28, 2021 @ 3:07 pm
Thank you, Michael. I was completely unaware of that. But when I listen to myself saying "I don't know", it has more than one possible tonal realisation, depending on which element takes the stress. Lacking any easy way to represent tone here, I will do my best —
I don't know (but someone else might) : H L M (high-low-middle)
I don't know (a confession of ignorance, with accompanying interest) L M (H-L).
I can't think of a situation in which I would want to stress don't.
[(myl) We need recordings!]
Jarek Weckwerth said,
August 28, 2021 @ 3:27 pm
I can tentatively confirm that Polish doesn't seem to have a separate contour for this. Nie wiem 'I don't know' can use more than one contour, but being just two syllables, I don't think any of them are distinguishable from the generic 'no' when done without real phonemic content (i.e. as [ʔmʔm]).
Robin Dawes said,
August 28, 2021 @ 3:47 pm
Philip Taylor writes "I can't think of a situation in which I would want to stress don't."
Occasionally the word "don't" is elongated and stressed to indicate irony:
"Why do I always have a headache on Saturday morning?"
"I do-o-o-on't know".
Antonio L. Banderas said,
August 28, 2021 @ 4:03 pm
Longman Pronunciation Dictionary (BrE ǁ AmE)
dəʊnt ǁ doʊnt —also, non-finally, especially before a consonant sound, dəʊn ‖ doʊn, as in ˌdon’t ˈknows.
This word has no weak form except occasionally də in don’t mind/know (cf. dunno də ˈnəʊ (ˌ)dʌ- ǁ -ˈnoʊ).
Antonio L. Banderas said,
August 28, 2021 @ 4:05 pm
See also
https://en.wiktionary.org/wiki/Talk:don%27t#Why_don't_you_%5Bt%CA%83%CA%8A%5D
Terry K. said,
August 28, 2021 @ 4:09 pm
While I can, like Philip Taylor, imagine different ways of saying "I don't know", nonetheless, I can also imagine the closed mouth "I don't know" being done one particular way.
Seems to me it's a version of a mumbled "I don't know", that tone and stress contour. So the versions with strong stress aren't actually relevant.
Michael Watts said,
August 28, 2021 @ 4:12 pm
For another point on the continuum of "I don't know" reduction, I knew someone who would usually spell it "iono".
David Marjanović said,
August 28, 2021 @ 4:16 pm
In German you symbolize shaking your head by saying [ʔḿ̩ʔm̩̀] – the only word with tones in the language happens to lack vowels.
(Nodding is symbolized by [ʔm̩ˈhm̩], which does not have fixed tones. There is no paralinguistic expression for "I don't know".)
There are tone languages in Africa and South America that can be, and often enough are, whistled, hummed or drummed: the tones carry enough information that the vowels and the consonants can be omitted if there's enough context.
[(myl) See "Pirahã Channels", 5/21/2006. See also Samuel Akinbo, "Representation of Yorùbá Tones by a Talking Drum: An Acoustic Analysis", 2019.]
Anthony said,
August 28, 2021 @ 4:32 pm
I think I just go up a third, then down a half-step. C E E-flat, for example.
JPL said,
August 28, 2021 @ 8:30 pm
Sorry, I can't give any recordings, but I'll try to indicate intonation contour by using numbers 1-4 for distinctive levels of pitch, 1 the lowest, 4 the highest, and people can reproduce them for themselves. "I don't know" seems to have several different contours indicating different presuppositions or discourse dynamics, but most (not all) would involve four tone levels or "notes", not three, due to "know" having rising or falling intonation. I just heard a three year old kid use the contour that the cartoon is probably referring to (reminding me of this post), but with words, not humming; but I think that it is only this one that (at least usually) gets the humming expression. That is the one with casual minor irritation that comes with a shrug and a "how should I know?" response: 3-4-2-3, hummed or with the words ("I don't know"), with rising intonation on "know". The emphatic version, that comes with the annoyed "how could you possibly think I know?" vibe would begin "know" with the lowest tone (1-2). The "I never thought of that" vibe would likely get a falling intonation: 2-3-4-1, but this would probably not get the hum. The "pondering" vibe might get 3-3-3-1. I suppose these could be hummed, but I think only the first case (casual irritation) is conventional. In Mende, a West African tone language, "I don't know" is "ngii koh" ("oh" standing in for the "open o"), with tones H-L, so not the same melody, and no possibility for further suprasegmental tones. One could probably say "ngii koh" with just the tones in a conversation (or with a "talking drum"), but I wouldn't say that it's conventional. Q: Is there a general unified semantic principle indicated by all cases of utterance-final rising intonation in English? (Another Q: Is the shrug universal?)
KMH said,
August 28, 2021 @ 9:53 pm
Years ago, Colbert had a segment on the Report where he claimed that some fraction of survey respondents had said "I don't know" (with a hand wobble and an intonation suggesting "I'd rather not say"), and another fraction had said "I don't know" (with a pinky to his lip and a coy intonation).
Jerry Packard said,
August 28, 2021 @ 10:07 pm
@Michael Watts
"I think of it as part of a set with uh-huh and uh-uh – they are the three possible answers to yes/no questions…"
In my dialect, you can makes the 3 notes with a closed mouth and 3 continuous nasal/voiced 'm's , which is how my son (32 yrs) does it. ‘uh-huh’ and ‘uh-uh’ require glottal stops: [ʔuh – huh] and [ʔuh – ʔuh] respectively.
Michael Watts said,
August 28, 2021 @ 10:25 pm
For me, uh-uh requires a glottal stop between the syllables, but you can have your mouth open or closed for any of the three answers. It doesn't matter. If I'm holding something between my teeth, I obviously won't be able to produce an [m], but I can still produce the "I don't know" contour that way.
I'm confused by everyone saying a glottal stop is required at the beginning of the first syllable. Word-initial glottal stops are not a normal feature of English. (Granted, these three words are already not normal, but I don't think that's one of the ways in which they're not normal.)
cliff arroyo said,
August 29, 2021 @ 3:45 am
"Nie wiem 'I don't know' can use more than one contour, but being just two syllables"
In informal daily usage 'nie wiem' is often slurred into something like 'nieem' or 'niem' though there is no particular special intonation contour that I'm aware of (it's my impression that intonation differences are one of the few surviving expressions of regional variation in the modern language.
David Marjanović said,
August 29, 2021 @ 3:52 am
No, but utterance-initial = postpausal ones are very common.
(Similarly, English doesn't have word-final devoicing, but it does have utterance-final = prepausal devoicing.)
Philip Taylor said,
August 29, 2021 @ 4:37 am
[(myl) We need recordings!] As requested.
J.W. Brewer said,
August 29, 2021 @ 6:43 am
I'm not entirely confident that my three-syllable-but-wordless "I don't know" has exactly the same pitch contour as when I say the actual words, not least because the former seems quite standardized and the latter might vary by context depending on emphasis, as has been noted above. I'm not technically competent enough to make and post a recording, but was trying to work it out in musical notation on a cheap toy piano we have in the house for the younger kids to bang on. The difficulty is that neither the (larger) drop from first syllable to second nor the (smaller) rise from second to third come out to an integral number of semitones. The initial drop is more than three semitones but less than four, while the final rise is less than one semitone. In my idiolect, at least; YMMV. (And no, this doesn't even mean that I end up precisely three semitones down from where I started.)
[(myl) Please send recordings — anonymity will be respected.]
Michael Watts said,
August 29, 2021 @ 8:52 am
Indeed, but that doesn't make them part of the word. I submit that uh-huh and uh-uh do not begin with glottal stops, and if you produce them without initial glottal stops, you will not even raise any eyebrows.
The fact that you're allowed to insert a glottal stop there is a separate consideration that doesn't even happen at the word level.
Andrew Usher said,
August 29, 2021 @ 9:42 am
I agree.
Initial glottal stops are _always_ sub-phonemic in English; their use or not means nothing and is normally not noticed. The only glottal stops required in English are in the middle of 'uh-uh', 'uh-huh', and some other paralinguistic expressions.
As for the "I don't know" sound (which doesn't feature glottal stops), just listen to the next post and you will surely get it.
k_over_hbarc at yahoo dot com
Michael Watts said,
August 29, 2021 @ 10:48 am
I believe American English also requires the reduction of /t/ to a glottal stop where it is followed by a more or less syllabic /n/ (as in "mitten" / "gotten" / "batten" / "hootenanny" etc.). You won't be misunderstood producing a [tʰ] there, but people will be aware that you're not pronouncing the word correctly, probably because you are foreign or are attempting to sound foreign. This is an odd case where intervocalic /t/ is systematically distinguished from intervocalic /d/.
Jerry Packard said,
August 29, 2021 @ 11:45 am
"If I'm holding something between my teeth, I obviously won't be able to produce an [m], but I can still produce the "I don't know" contour that way."
I have no problem producing an [m] while holding something between my teeth, as long as my lips are closed.
And yes, initial glottal stops are _always_ sub-phonemic in English, but they are virtually always there, preceding an 'initial' vowel (try to produce an initial vowel without a preceding glottal stop).
My first phonology teacher, Linda Waugh, posited initial glottal stops to be contrastive in some English dialects, contrasting with initial [hw] (vs. [w]).
Jerry Packard said,
August 29, 2021 @ 11:57 am
"I believe American English also requires the reduction of /t/ to a glottal stop where it is followed by a more or less syllabic /n/ (as in "mitten" / "gotten" / "batten" / "hootenanny" etc.). "
Also when followed by 'syllabic' /l/. But most dialects also require that /t/ to occur, albeit in unreleased form. Contrast standard 'bottle' [boʔt ̚ l] with
Brooklynese [boʔ l].
Jerry Friedman said,
August 29, 2021 @ 12:02 pm
This is very familiar to me, and is usually accompanied with a shrug. I may be able to send a recording.
I've seen a complaint about teenagers who never open their mouths to adults. They speak only in response to questions, and the only possible answers are "m-hm", "m-m", and "m-m-m", for "Yes", "No", and "I don't know".
I left out the glottal stops. In my English, a glottal stop is required in the negative "m-m" and "uh-uh". Without it, the latter might be written "huh-uh" (which is not in my idiolect), though the meaning would be the same, which I suppose makes it sub-phonemic.
Am I right in thinking that for some or all RP speakers, a glottal stop is required in every word with a word-initial vowel? "ʔIt was ʔabsolutely ʔawful."
(Linguistic adjective-final compounds such as "word-initial" always reminds me of botanical ones such as "white-hairy" 'covered with white hairs', though I don't think they're the same thing. The linguistic ones start with a noun and the botanical ones start with an adjective—most of the time?)
Philip Taylor said,
August 29, 2021 @ 12:24 pm
Jerry — « Am I right in thinking that for some or all RP speakers, a glottal stop is required in every word with a word-initial vowel? "ʔIt was ʔabsolutely ʔawful." ». Certainly not for all, and I am very doubtful about "some" — to those speakers of RP with whom I am familiar, the glottal stop is complete anathema, and most would do everything in their power to avoid producing one. High RP, that is —it may well feature in what Jack Windsor Lewis would term "General British".
Philip Taylor said,
August 29, 2021 @ 12:37 pm
As far as I can see, the only terms in the LPD with an initial glottal stop are "ahem", "uh oh" and "uh uh". The IPA transcription of "ahem" does not accord with the recorded speech, however — the given transcription is [ʔm ʔmː], whilst the recorded speech sounds to me more like [ʔʌ hʊm].
Michael Watts said,
August 29, 2021 @ 12:56 pm
I don't agree – the glottal stop is not standardly required, or present, before /l/, and as you note even if it were present it would not be reduced from the /t/, since the /t/ is still there. The standard is a [t] or [ɾ] with lateral plosion, and there is no distinction between /t/ and /d/ in this context.
That is to say, the treatment applied to /d/ (but not /t/) when before /n/ is applied to both /t/ and /d/ when before /l/.
Jerry Packard said,
August 29, 2021 @ 3:00 pm
« Am I right in thinking that for some or all RP speakers, a glottal stop is required in every word with a word-initial vowel? "ʔIt was ʔabsolutely ʔawful." »
I think so, unless the V is a glide like 'y' or 'w'. I don't know about 'required', but I think it does occur there. Try to say "It was absolutely awful" without saying "ʔIt was ʔabsolutely ʔawful". Charles Hockett (my first phonetics teacher) called it a 'glottal catch' rather than 'glottal stop' in that position.
"That is to say, the treatment applied to /d/ (but not /t/) when before /n/ is applied to both /t/ and /d/ when before /l/."
Yes, I think you may be right.
MM said,
August 29, 2021 @ 7:05 pm
To my ear this usage implies the respondent finds the question unanswerable or not worth answering. It's the response I give when my kid asks how to craft some obscure object in Minecraft. It probably isn't the answer to the question: 'Is the baby strapped into the travel capsule?'
And here's one in the wild from 1994:
https://www.youtube.com/watch?v=Tk4yyqXi8Xc
[(myl) The audio and f0 track from that version:
More on that (and other versions) later…
]
Josh R. said,
August 29, 2021 @ 7:31 pm
Michael Watts said:
"I believe American English also requires the reduction of /t/ to a glottal stop where it is followed by a more or less syllabic /n/ (as in "mitten" / "gotten" / "batten" / "hootenanny" etc.). "
This was the cause of a personal linguistic existential crisis when it came to modeling English pronunciation for my toddler daughter. As we live in Japan, I am her primary source of English (and interlocutor), so I began to be hyper-aware of my own pronunciation, and any peculiarities I might pass down to her. I've long been aware of the American intervocalic /d/ for /t/ (e.g., water), but I'd never been as clearly cognizant of glottal stops before /n/. I was telling her to button her pajamas when I started thinking to myself, "Wait, 'buh'in'? Is that right? Is that dialect? Am I just pronouncing it lazily?"
20 years in another country can really give your linguistic self-confidence a hit.
Jarek Weckwerth said,
August 30, 2021 @ 2:52 am
@Philip Taylor: "terms in the LPD with an initial glottal stop"
As people have said in this thread, the glottal stop is subphonemic, and therefore not normally shown in dictionary transcriptions.
The initial glottal stop is also called "hard attack" by some people.
David Marjanović said,
August 30, 2021 @ 3:34 am
Now, I haven't been listening that much to the last few prime ministers, but this strikes me as pauses for emphasis: "It was — absolutely — awful." The glottal stops are then just postpausal.
There are several lines of evidence that Old English did have obligatory prevocalic glottal stops preceding either words or maybe stressed syllables*, and a few that Middle English lacked them, from alliteration and syllable mergers in poetry.
* The latter is the case in northern German today.
Benjamin E. Orsatti said,
August 30, 2021 @ 7:25 am
…and then there's the Pittsburghese "emphatic no", pronounced: "næ, huh(glottal stop)-uh".
Andrew Usher said,
August 30, 2021 @ 7:44 am
I did not mention the American use of glottal stops before syllabic 'n' to simplify, because they're not actually _required_ to speak properly, and there are more varieties of English than just American. Indeed, I've heard that some Americans (where?) prefer the flap rather than the glottal stop in words like 'mitten'.
But now I'll give a more detailed summary of that: when /t/ precedes /n/ (by definition unstressed), it is normally pronounced as /?/. However, when the segment preceding /t/ is /n/ or /l/, a full [tʰ] is at least an acceptable variant, as in 'fountain' or 'Clinton'. When the segment preceding is /r/, we normally get a flap [ɾ] though [?] is not impossible (note that 'kindergarten' is said as if spelled 'kindergarden' and thus not a real exception). Finally. when it is a vowel, the glottal stop is definitely standard.
But in the most careful (American) speech, I think you get an unaspirated [t] in the last two cases. Certainly when 'demonstrating' them, I feel my mouth making the /t/ closure, and then the difference between 'mitten' and 'midden' becomes one of voicing only.
Richard Hershberger said,
August 30, 2021 @ 11:11 am
Paying amateurs for ideas: Here's my idea for a mystery novel. The detective realizes the supposed Brit is actually an American by tossing an object at him, which he catches with his left hand rather than with both. You are welcome. I accept cash or checks/cheques.