« previous post | next post »

Yesterday's post "A stick with which to beat other women with" discussed the duplication of prepositions in the title phrase, and a commenter complained that

The woman interviewed has a pretty mediocre command of English (she doesn't pronounce a single coherent sentence and keeps stuttering) although she is an actress speaking in her native language. That she would make mistakes in her own language is thus regrettable but not especially surprising. I am not unaware that the concept "mistake" does not enjoy stellar prestige among linguists, but why is that particular error worthy of a blog entry?

As another commenter observed, my original post used the phrase "performance error" to describe the possibility that Emma Watson's preposition doubling was a mistake rather than a bona fide syntactic variant.

But my point today is that verbatim transcripts of spontaneous speech are often full of filled pauses, self-corrections, and other things that must be edited out in order to create what that commenter would count as a "coherent sentence". And this is true even for people who have risen far in the world on the basis of their ability to impress others in spontaneous verbal interaction.

For example, here's Justice Anthony Kennedy's first question during the 11/8/2016 oral argument in BANK OF AMERICA CORPORATION, ET AL., Petitioners v. CITY OF MIAMI, FLORIDA, Respondent:

Your- is your concession- um
oh not a concession your- y- your formulation
that the city can sue sometimes
are you thinking that the city might be
in the same position as Home was
in um
was it the- the- uh Ha- the Havens case?

Or this passage from remarks by Stephen K. Bannon in a panel about "The Future of Conservatism", held at the National Press Club in Washington DC on 9/24/2013, and previously discussed in "The Narrow End of the Funnel", 8/18/2016:

w- I think it's given us a- uh I think it's given us a voice, I mean I think you don't have to go through the mainstream media
filter any more you don't have to even real-
concern about uh
you know television or- or- or cable or whatever I mean we have a story up today of people that came to us from a-
the Liberty Institute
on this Craig James situation where Craig James was a broadcaster who was basically fired
uh because uh in a- in a senatorial campaign
he had uh discussed uh traditional marriage
uh and after he was already working there somebody found the tape and they said
you can't work here any more
that- the- the- in- the- today this thing's gone viral
uh it's a story we worked on it does- didn't have any
TV backing any uh
any uh
broadcast backing or cable backing and we've gotten it up and it's already you know is going to have a million page views today and
and- and go throughout so

These are random examples from the first couple of transcribed recordings that I looked at — I could add examples from presidents, philosophers, scientists, novelists, poets, talk show hosts, and even usage experts.

The point: this is what (most) spontaneous speech is like. Emma Watson is no more disfluent or incoherent than Anthony Kennedy or Stephen Bannon or any of the rest of them. People who speak in perfect written paragraphs are almost pathologically unusual.

(See also "The rhetorical style of spontaneous speech", 8/16/2016.)




  1. David L said,

    March 8, 2017 @ 10:48 am

    I did a little bit of live radio years ago and, if I can be immodest, I was pretty good at it. I mean that I was able to speak reasonably fluently and coherently without a lot of forethought. I have noticed that some people who are evidently pretty smart are not able to do that, apparently because they have to rehearse what they are about to say before they say it, and thus pause and stutter and change direction. Whereas I was able to make stuff up as I went along and still have it come out in moderately fluent fashion.

    Listen to sports commentators, for instance. The best of them of them can keep talking (and talking and talking…) with little hesitation or stumbling. Or by contrast, think of some academic teachers, no doubt very bright, who were terrible lecturers.

    Speaking extemporaneously is an ability in its own right, is what I'm saying, with not much connection to intelligence (myself excepted, of course).

    [(myl) One issue — ripe for study, I think — is the difference between what we might call fluent and non-fluent disfluencies. Some people manage to give the impression of a clear and eloquent presentation despite many filled pauses and self-corrections; in other cases a similar frequency of editable events seems entirely incoherent. Many well-known broadcast personalities are, so to speak, fluently disfluent.

    As I've observed a number of times, one of the most striking characteristics of Donald Trump's spontaneous oratory is the almost total lack of filled pauses and self-corrections.]

  2. Rube said,

    March 8, 2017 @ 11:02 am

    @ David L: come to think of it, wasn't J.R.R. Tolkien, who I suppose could be regarded as pretty fluent in his own language, and God knows how many others, a famously terrible lecturer (except possibly when he was reciting Beowulf)?

  3. David L said,

    March 8, 2017 @ 11:53 am

    @myl: what you call 'fluent disfluency' is what I would describe using the less scientific term 'winging it.' In other words, when you start to lose track of what you're saying or where you're headed, instead of going 'um, ah, I mean, like' you somehow manage to keep the speech motor running while some other part of your brain is figuring out a rescue plan.

    Stand-up comedians would be interesting to study, perhaps. A lot of what they say is rehearsed, obviously, but the good ones come up with a lot of ad libbed banter (to use another scientific term) as they interact with and riff off the audience reaction.

    [(myl) In this post I give an audio clip and verbatim transcript from a performance by Louis CK…]

  4. Ross Presser said,

    March 8, 2017 @ 12:04 pm

    People who speak in perfect written paragraphs are almost pathologically unusual.

    Or more commonly, reading from a prepared, written speech.

  5. Chips Mackinolty said,

    March 8, 2017 @ 12:17 pm

    As a former working print journalist, there is a sense that we have always falsified transcripts of interviews, in the following ways:
    1 We remove the "ums" and "ers", let alone the partial words and self-corrections. It would be too hideous to transcribe, from notes, shorthand or tape, let alone for someone to read. [And, without doubt, our interrogations of the interviewee are equally disfluent/incoherent.]
    2 We edit responses in a variety of other ways, from the point of view of succinctness and word count, and this leads inevitably to short sentences being reported rather than the endless sentences/stream of consciousness most of us speak in everyday speech. And we certainly edit out the interviewee calling us by name.
    3 We edit to "fit" the yarn, removing irrelevancies and diversions
    4 And to avoid endless (sic) denotations, we are inclined to amend simple grammatical errors or non-formal usages [eg ask vs aks].

    Which is why interviews in "print" generally look formal, coherent and fluent. A good deal of that is the pressure of word count: if you are asked to file 750 words, the last thing you want to do is waste the precious space, let alone face the wrath of news editors and sub-editors hounding you for filling your copy with gratuitous crap.

    It's obviously different for broadcast journalists–there's only so much you can get away with editing audio and/or vision.

    But there are times when, in print and broadcast, the temptation to "let it run as it was said" is irresistible. It's happening with Trump now, and back in the early 1970s it was done to the then Prime Minister of Australia, Billy McMahon. Journalists of the time took delight in reporting him verbatim as a vivid demonstration of how inept he was in making much sense about anything.

  6. Jerry Friedman said,

    March 8, 2017 @ 12:58 pm

    People who speak in perfect written paragraphs are almost pathologically unusual.

    That's an interesting way to put it. Some years after having lunch with you, I have the impression that your spontaneous speech was very low in disfluencies (maybe especially low in the non-fluent kind), though I couldn't say whether it consisted of perfect written paragraphs.

    David L.: Or by contrast, think of some academic teachers, no doubt very bright, who were terrible lecturers.

    There are many ways to be a bad or even terrible lecturer other than disfluency. I've perpetrated some myself—I hope not too often and with decreasing frequency.

  7. David Spencer said,

    March 8, 2017 @ 1:39 pm

    John Rawls was arguably the most brilliant moral philosopher of the 20th century. He was also arguably the worst speaker of all time.

  8. D.O. said,

    March 8, 2017 @ 2:12 pm

    We fail to capture so much of the speech when transcribing it (pauses, speed, intonation, rhythm, changes in volume, gesticulation, and a lot of other stuff that I totally fail to remember) that it would be bordering on dishonest to represent the transcript full of disfluencies as "real speech" or "what was really said".

    [(myl) Yes, that was the point of the cited post "The narrow end of the funnel", 8/18/2016.]

    We may also think that speech (oral or written) is a two-agent communication, even if there are no replies, and perception mechanism by a reader might be very different from the perception mechanism of the listener. It means that transcribed unedited speech does a disservice to both the speaker (and unwitting writer) and reader (and missed-out listener).

  9. D.O. said,

    March 8, 2017 @ 2:38 pm

    BTW, as speech recognition technology becomes more and more common, the "recognition" part should become more and more adept at weeding out the disfluencies. OTOH, on "speech" part, people doing dictations to automatic transcription programs probably came up with a few tricks how to speak relatively fast and freely and, at the same time, not to do too much editing.

  10. Kevin said,

    March 8, 2017 @ 3:23 pm

    Or how about Shakespeare:

    "That fair for which love groan'd for and would die,
    With tender Juliet match'd, is now not fair."
    (Romeo and Juliet, Act 2, Prologue)

  11. Geoff said,

    March 8, 2017 @ 4:11 pm

    I suspect many people are quite unaware of how disfluent most normal speech is, since:
    – in speech we're so used to ignoring it
    – they've never come face to face with a real verbatim transcript.

    In the organisation where I work making edited transcripts of speech, think of the job as, by time spent: one third putting it down verbatim using voice recognition; two thirds fixing it up to make it readable and not sounding stupid. I suspect many people would be surprised that those are the proportions.

    I can foresee the computer getting smarter and smarter at putting down the verbatim transcript to avoid eggcorns and mondegreens (using context to distinguish plural S from possessive S etc). I can't imagine the computer any time soon replacing the editing task, which often requires interpreting many lines together and adjusting dependencies at some distance.

    And yes, the contrast between fluent disfluent speakers and non-fluent disfluent speakers is interesting.

  12. Rubrick said,

    March 8, 2017 @ 7:05 pm

    I'm unsurprised to learn (from Jerry Friedman's comment) that MYL is exceptionally fluent in spontaneous speech.

    I believe John Updike was one of those "pathological" cases who came pretty close to speaking in perfect prose, although I'll admit I've only heard him in onstage interviews, not casual hallway conversations.

  13. Ray said,

    March 8, 2017 @ 7:14 pm


  14. David Morris said,

    March 9, 2017 @ 2:40 am

    I sometimes wonder how I sound to my students.

  15. Monte Davis said,

    March 9, 2017 @ 6:27 am

    Hearty agreement with Mark, Chips and Geoff. After decades as a magazine interviews editor and then a speechwriter working from my taped & transcribed input sessions, I'm *still* surprised at the level of disfluency on paper that I simply did not hear. Most usage Nazis — entirely aside from the ideological, cultural, and class axes they're usually grinding — seem never to have reflected on the grossly obvious fact that printed speech *sits there* for a level of analysis we can't, don't, and shouldn't apply aurally.

  16. Lane said,

    March 9, 2017 @ 7:43 am

    I have also had lunch with Mark, and found myself writing later "He speaks so slowly that when I listen to my tape and play it on fast speed, I sound like an auctioneer and he sounds normal."

  17. Edwin Schmitt said,

    March 9, 2017 @ 8:06 am

    Perhaps part of the problem is that during interviews and presentations one is expected to provide a great deal of information in a single stream. This takes the intersubjective aspects of language out of the picture. Frankly, I find both interviews and presentations to be completely awkward forms of communication and often question their effectiveness in transmitting information. After all the oratory structure of human language was historically reserved for the re-telling of myth (which is certainly about much more than just the transfer of information from one person to another), but even then there are cultures which require a back and forth in order for a myth to be told in its totality. I think a fascinating study would be a genealogy of the interview or presentation form of language and the way such forms of communication dictate the kind of information that is transmitted, perhaps in contrast to the kind of information that is transmitted through general conversation.

  18. Rodger C said,

    March 9, 2017 @ 8:28 am

    Obligatory geezer reference: When the Nixon tape transcripts were first published, disfluencies and all, many people said he sounded like a blithering idiot, but when the tapes themselves are listened to, it's obvious he's just talking.

  19. Jan Schreuder said,

    March 9, 2017 @ 9:05 am

    Gore Vidal was, famously or notoriously, fluent when speaking, especially about his many enemies.

  20. Jonfrum said,

    March 9, 2017 @ 1:03 pm

    The fields of both speech pathology and linguistics have made a study of exactly this type of error of speech. They are perfectly normal, and have various possible explanations. Here, most likely, we have a pausing device used to search for the most appropriate finish to the sentence. Nothing to see here – and I'd love to hear a recording of spontaneous speech for anyone who disagrees.

    Regarding the commenter quoted above – he (?) seems to be using 'stuttering' in a disparaging sense. In other words, she's a dumb-ass, who can't speak clearly, so she stutters. As a stutterer – that is, a person who suffers from a neurological developmental disorder, I don't appreciate that connotation being applied to my particular pathology. Or am I just being a retard?

    [(myl) It's clear that "stuttering" (or "stammering") is a very different thing from self-corrections of the kind exhibited in Emma Watson's interview. The behavior is different, and I'd be happy to wager that the underlying physiological motor control patterns are quite different as well.]

  21. Charles Antaki said,

    March 12, 2017 @ 4:55 am

    Just to add a performative penn'orth – latent in some posts is the reminder that Watson was in a dialogue, not giving a speech; some at least of her disfluencies are readable as, and meant to be readable as, displays of her dialogic point, viz., upset-to-the-point-of-near-incoherence. If you struggle to express idea X to an enquiring interlocutor and potentially hostile others, you're indexing how odd and alien idea X is.

RSS feed for comments on this post