More on UM and UH
« previous post | next post »
A few days ago ("Fillers: Autism, gender, and age" 7/30/2014), I noted an apparent similarity between male/female differences in UM/UH usage, and an autistic/typical difference reported in a poster by Gorman et al. at the IMFAR 2014 conference.
This morning I thought I'd take a closer look at the patterns in a large published conversational-speech dataset. Executive summary:
- There is a large sex difference in filled-pause usage, favoring males by about 38%
- There is an enormous sex difference in UM/UH ratio, favoring females by about 310%
- These sex differences are mainly driven by the difference in UH usage, which favors males by about 250%
- Older speakers use UH more and UM less, resulting in a large decrease of UM/UH ratios
The general pattern of gendered filled-pause usage in English has been at least partly replicated in several other datasets, including the spoken part of the British National Corpus, but the details are sometimes quite different. (See my earlier post, and planned future posts, for some discussion.) But all the important questions remain open, for example:
- Are the sex effects due to functional, iconic, or physiological differences between UM and UH, or are they arbitrary gender markers?
- Do the age effects reflect a change in progress, or a life-cycle effect (e.g. due to changes in sex hormone levels)?
- Are the patterns the same or different across geographical, socio-economic, and ethnic varieties of English?
- Are there analogous phenomena in other languages?
I'm looking at data from the Fisher collections (part 1 and part 2). If we lump all the 11,972 speakers together into categories based on the gender assessments made during call auditing, we get these overall statistics for the two types of filled pause:
| Total words | um | um % | uh | uh % | um+uh % | um/uh ratio | |
| Males | 10,163,101 | 82,659 | 0.81% | 120,563 | 1.19% | 2.00% | 0.686 |
| Females | 12,986,418 | 128,368 | 0.99% | 60,395 | 0.47% | 1.45% | 2.125 |
These numbers establish the basic pattern for this dataset:
- females overall use um somewhat more than males (about 22% more on average);
- males use uh a lot more than females (about 250% more on average);
- males use filled pauses (um+uh) more than females (about 38% more on average);
- the um/uh ratio is more than 3 times greater for females than for males.
The effects are mainly driven by the large difference in uh usage.
There's a modest amount of accommodation to interlocutor sex — that is, males use uh about 14% less often when talking with a female rather than a male, and females use uh about about 20% more often when talking with a male rather than a female:
| Male Interlocutor | Female Interlocutor | |||||
| Total words | uh count | uh % | Total words | uh count | uh % | |
| Male Speaker | 6322608 | 78680 | 1.24% | 3840493 | 41883 | 1.09% |
| Female Speaker | 3656054 | 19433 | 0.53% | 9330364 | 40962 | 0.44% |
There's a smaller accommodation effect for um — males use um about 8% more often when talking with a female, and females use um about 1% less often when talking with a male:
| Male Interlocutor | Female Interlocutor | |||||
| Total words | um count | um % | Total words | um count | um % | |
| Male Speaker | 6322608 | 49973 | 0.79% | 3840493 | 32686 | 0.85% |
| Female Speaker | 3656054 | 35922 | 0.98% | 9330364 | 92446 | 0.99% |
As a result, overall male um/uh ratio shifts by about 23% in the female direction with an opposite-sex interlocutor, while female um/uh ratio shifts by about 22% in the male direction:
| Male Interlocutor | Female Interlocutor | |
| Male Speaker | 0.635 | 0.780 |
| Female Speaker | 1.849 | 2.257 |
At the level of individual speakers, the basic patterns hold up. For this part of the study, I restricted the dataset to the 19,906 calls where the evaluation by post-call auditors agrees with the subject's demographic information collected before the call — in a small fraction of cases, it seems that a different member of the same household answered the phone and participated in a call.
Males tend to use filled pauses (UM+UH) somewhat more than females do, as shown in the "violin plot" below:
Much of the difference is in the high-usage tail, so that the ratio of median UM+UH percentages (male 1.59/female 1.29 = ratio of 1.23) is quite a bit smaller that the ratio of means (male 2.00/female 1.45 = ratio of 1.38).
Breaking it down by specific fillers, we see again that the biggest difference is in UH usage, and especially in a long tail of males with especially high UHpercentages:
| Median UM% | Median UH% | |
| Male | 0.659% | 0.701% |
| Female | 0.889% | 0.246% |
For looking at individual variation in um vs. uh usage, it makes more sense to use the proportion of um in the total — UM/(UM+UH) — rather than the UM/UH ratio, which is unstable for small samples. If we limit consideration to those speakers whose transcripts contain at least 10 filled pauses, we get this:
The median UM/(UM+UH) proportion for the males is 0.500, and for the females 0.786 (57% greater).
The violin plot for individual variation in UM/UH ratio is dominated by its long tail on the high side:
The median UM/UH ratio for the males is 0.719, and for the females 2.63 (366% greater).
In order to look at age effects, I've divided speakers at the 33rd and 67th percentiles by age, which in this collections comes out to those younger than 28, those between 28 and 40, and those older than 40.
Here are the relevant plots for UM% and UH%:
 |
 |
And the corresponding numbers:
| UM% | UH% | |
| Male <28 | 0.90% | 0.85% |
| Male 28-40 | 0.87% | 0.93% |
| Male >40 | 0.77% | 1.36% |
| Female <28 | 1.08% | 0.28% |
| Female 28-40 | 1.05% | 0.35% |
| Female >40 | 1.01% | 0.58% |
The result, obviously, is a decline in UM/UH ratio with age:
The numbers (UM/UH ratios):
| Age <28 | Age 28-40 | Age >40 | |
| Males | 1.064 | 0.931 | 0.566 |
| Females | 3.851 | 2.978 | 1.738 |
And a log plot to suggest the approximately proportional change with time:
Bathrobe said,
August 3, 2014 @ 9:15 am
Quite personally, I find 'um' and 'uh' somewhat different. 'Um' is the sound you make when you're actively 'wondering' or trying to recall something. 'Uh' is the sound you make when you're trying to think of the next word. 'Um' therefore sounds less confident and more tentative. My own feeling is that 'um' would perhaps sound more appropriate for a younger or female demographic, possibly since they would be less concerned about sounding authoritative. For a male, especially an older male, 'um' would sound too diffident and lacking in confidence or authority to be used too often.
S said,
August 3, 2014 @ 11:07 am
This is tangential, but is there also data coded for sexuality? It'd be interesting to see if there's also a difference in the ratio between gay and straight speakers, which I suspect there would be if there is Something Going On.
[(myl) Good point. But that's not one of the demographic variables available for this dataset.]
bratschegirl said,
August 3, 2014 @ 11:36 am
I'm surprised to see that men use filled pauses more than women; I would have expected the opposite. For either, though, I wonder whether it has to do with "I'm still making noise here and therefore haven't given up the floor," whereas an unfilled pause would encourage someone else to begin speaking? Sort of a vocal marking of territory.
[(myl) Could be. There certainly seem to be functionally-different uses of filled pauses — but I suspect it would be hard to code them in an inter-subjectively stable way. Maybe not impossible, though.]
D.O. said,
August 3, 2014 @ 5:40 pm
I decided to look at distribution of uh and um inside the conversation turns. But I don't have access to the Fisher dataset, only to 2 examples from parts 1 and 2 — a drop in the ocean comapred to the dataset with which Prof. Liberman works. Both sides in both conversations are apparently women. There are the total of 582 conversation sides separated by pauses, not necessarily by turns (i.e. the same speaker sometimes continued the next side). There were 13 uh's and 17 um's. Here's the distribution.
Uh (13 total): 6 side initials, 5 of them after the turn; 4 side finals, 2 of them before the turn; 3 middle of the side.
Um (17 total): 4 side initials, 1 after the turn; 9 finals, 3 after the turns; 4 middles
From this tiny sample, it seems that uh's have larger propensity to be near the side turn (7/13) than um's (4/17). Make of it what you may.
Carl said,
August 3, 2014 @ 7:10 pm
Like most sex differences, this one is clearly traced to our evolutionary background. I the wild jungle, our ancestors needed to use um and uh to protect themselves from roaming tigers and hippopotamuses. Men would say UH to repell the tigers of the savannah and women would say UM to alert others to hippopotamuses in the cave. Attempts to pin this difference on culture are just so much more PC, feminist, hysterical claptrap.
Bathrobe said,
August 3, 2014 @ 8:00 pm
@ Carl
I find this the most convincing explanation yet. Do you happen to know if it applies only to English-speaking cavemen, or does it apply to all language families?
Rubrick said,
August 4, 2014 @ 2:39 am
I must disagree with Carl. To me it seems clear that the difference comes down to women's more acute sense of smell. This makes them more self conscious about whether their own breath might smell foul, and so they are inclined towards the close-mouthed "ummmm" rather than the open-mouthed "uhhhhh".
Nate said,
August 4, 2014 @ 3:16 am
"Are the patterns the same or different across geographical, socio-economic, and ethnic varieties of English?"
Do all varieties even use "uh" and "uhm"? No, they don't.
Larry said,
August 4, 2014 @ 8:52 am
I suspect men use /uh/ as a place-holder. Women use /um/ as a backchannel indicating, "that's interesting, I'm thinking about it."
rosie said,
August 4, 2014 @ 9:24 am
@Nate Quite. There's also "errrr……" which is useful when you need time to work out what to say next, and don't want the other person to interrupt.
Anthony said,
August 4, 2014 @ 9:41 am
Isn't "er" just the British way of writing "uh"?
Brett said,
August 4, 2014 @ 9:57 am
@Anthony: I have encountered people who used a rhotic "er" as an actual interjection. I assume they picked it up from reading British (or similarly non-rhotic). There's nothing unusual about picking up a word from written sources and mispronouncing it, but it's seems odder when it's an interjection.
Matt_M said,
August 4, 2014 @ 10:34 am
@Anthony: I'm not so sure. I think the difference between written "uh" and "er" is actually meaningful to many non-rhotic speakers. To me (a speaker of Australian English), "uh" and "er" are both (of course) non-rhotic, but "uh" suggests an open (central) vowel, while "er" suggests a mid-central vowel. I think I make use of both of these vowels in making hesitation interjections.
Kyle Gorman said,
August 4, 2014 @ 11:15 am
According to the OED, "er" and "erm" are orthographic variants of "uh" and "um", and phonetically identical thereto.
Matt_M said,
August 4, 2014 @ 11:23 am
@Kyle: but why would "erm" and "um" be phonetically identical when "berm" and "bum" are pronounced so differently by both rhotic and non-rhotic speakers of English? I don't think many people look up the pronunciations of "uh" and "er" in the dictionary before writing them.
Kyle Gorman said,
August 4, 2014 @ 11:34 am
@Matt_M the assumption that English words which share the last three letters should be pronounced with the same rime is obviously false.
I could similarly ask "but why would the OED make a false report?" Perhaps you're right, but it's acoustically awfully subtle
(similarly, people used to think that non-rhotic speakers in NYC pronounced "source" and "sauce" the same until Labov, Yaeger, and Steiner 1972 did an acoustic analysis), or, perhaps this equivalency is true in prestige variants the OED describes, but not in some other variants of English.
Matt_M said,
August 5, 2014 @ 6:54 am
@Kyle: all good points, but:
I can't think of any other words in English where the spellings "er" and "u" have the same sound;
The difference in vowel quality beween, say, "lurk" and "luck" in Australian and Southern British English is not subtle at all. The difference in vowel height is at least as great as between "bad" and "bed", for starters, and there's also some light lip rounding on the "er" vowel;
Although I'd normally trust the OED, I wonder what kind of methodology they'd employ in this kind of case to determine the pronunciation? These kinds of words don't have canonical pronunciations in the same way as, for example, "police" does (I can hardly imagine a teacher correcting a student's pronunciation of "uh" or "um"). I imagine that every person who comes across "uh", "er", "um", or "erm" in print forms their own impression of how it's pronounced, without much opportunity for feedback from others. And the idea that they represent different vowels is certainly a plausible one for a non-rhotic speaker — so I imagine that lots (perhaps a majority?) of such speakers actually do think of "er" and "uh" as representing different vowels.
ajay said,
August 5, 2014 @ 9:28 am
I imagine that every person who comes across "uh", "er", "um", or "erm" in print forms their own impression of how it's pronounced, without much opportunity for feedback from others. And the idea that they represent different vowels is certainly a plausible one for a non-rhotic speaker — so I imagine that lots (perhaps a majority?) of such speakers actually do think of "er" and "uh" as representing different vowels.
I may be misreading this (=Dutch: he is not confident about reading my English; British: you are clearly insane) but are you saying that people start filling pauses with "er" and "uh" only after encountering the words in print and guessing for themselves how to pronounce them? That doesn't seem likely.
The point has been made that UM is basically UH with the mouth closed, and I can't help thinking this might be important; women more aware of the unsightly effect of the mouth hanging open during a pause in speech, men just not caring?
UH is also louder, maybe. Perhaps this is an anti-interruption technique used by men, explaining why they don't get interrupted as often. If you finish a sentence and say UM, people might not hear you, and will therefore start talking…
Matt_M said,
August 5, 2014 @ 9:54 am
@Ajay: "are you saying that people start filling pauses with "er" and "uh" only after encountering the words in print and guessing for themselves how to pronounce them? That doesn't seem likely."…
No, that's not what I'm saying. What I'm trying to suggest is that people produce and hear a variety of more-or-less central vowels as hesitation interjections, and that when non-rhotic speakers encounter "er" and "erm" in print, they may well be inclined to match these spellings to an interjection (that they're already familiar with) with a mid-to-high vowel, and when they encounter "uh" and "um" in print, they may well be inclined to match these spellings to an (already familiar) interjection with an open vowel.
Bathrobe said,
August 5, 2014 @ 7:55 pm
@Matt-M I think the difference between written "uh" and "er" is actually meaningful to many non-rhotic speakers.
In most cases it probably is, but with the indeterminate vowel in 'uh' or 'er', I wonder if there is any significance in the difference. It could be that in trying to find a way to represent this sound, people have simply alighted on different spellings that seem somehow 'near enough'. I mean, if you had to write that sound and didn't have formal orthography to guide you, what would you write? A rhotic speaker might write 'uh' (avoiding 'ah' because that already has a different connotation); a non-rhotic speaker might write 'er' in order to capture the sound. This is not necessarily schwa. For example, the final syllable in 'mother' in my non-rhotic speech is exactly the sound I would use for 'uh' — and it's not schwa! (Come to think of it, I probably wouldn't use schwa for 'uh'; it sounds unnatural.) I think you're possibly reading too much into the different spellings.
That rhotic speakers might pick 'er' up from written sources and use it in their speech isn't unusual, not much different from people who say 'tut tut'.
The Art of ‘Um’ and ‘Uh’: Different Vocal Pokes for Different Media Folks - PRNewser said,
August 12, 2014 @ 1:15 pm
[…] noted on his blog, Language Log, Liberman uncovered some telling numbers regarding these vocal crutches that shouldn't […]
RachelB said,
August 12, 2014 @ 5:45 pm
Anecdotally speaking, my partner (who is male) has observed that coming from me, "um" indicates that I am in the process of coming up with a polite way of disagreeing with what someone just said, or declining an offer someone has just made.
Writing and speaking are not the same thing, but I see variations on "um, [expression of disagreement]" much more frequently in writing than "um, [expression of agreement]."
Has anyone else observed, either in speech or writing, a strong correlation between "um" and expressions of disagreement?