The over/under on linguistic discovery
« previous post | next post »
Geoff Pullum, "The world's greatest grammarian", Chronicle of Higher Education 4/3/2017 [emphasis added]:
We mostly did 11-hour days, starting as soon after 7 a.m. as we could and working till 6 p.m., breaking for a short lunch at 1 p.m. to discuss the morning’s work. Virtually every day we would find over our sandwiches that we had discovered something new about English syntax that no one had never known before. Far from being a period of tedious recording of well-documented facts about the world’s best-documented language, it was actually the most exciting research period of my life.
In a comment on a post yesterday ("Blasphemous", 4/4/2017), someone remarked:
At the risk of being blasphemous myself, isn't this a misnegation from Geoff Pullum in his Lingua Franca submission today about Rodney Huddleston?
I suspect that it's a typographical or editing error, unless it's an example of Geoff's often-subtle humor.
But my reaction to that passage was to be surprised that it took them until lunch time to come across something new in English syntax. In examining an arbitrary spoken passage, even in well-studied languages like English, the over/under for observing a new (and interesting) linguistic phenomenon is about 10 seconds. Or so I commonly assert to students.
Wikipedia explains that:
An over–under or over/under (O/U) bet is a wager in which a sportsbook will predict a number for a statistic in a given game (usually the combined score of the two teams), and bettors wager that the actual number in the game will be either higher or lower than that number.
So the wager here is that if you examine a random stretch of English speech closely, within 10 seconds (of speech time, not your analysis time) you'll notice an instance of an interesting general pattern that has never been systematically studied in the linguistic literature.
As an illustration, here's a short clip from the recently-released podcast S-Town (Chapter 07, starting about 19:57), which I pulled out as an example of a certain prosodic style:
oh it's wonderful at homecoming to go to the turnip green supper
and there's a bonfire and everybody you went to school with and
everybody brings a dish I mean and I'm sure
you'd know who mi- you'd know that Miss Layler's made the turnip greens and
you know to oh try her coconut pie, Miss Daley's banana pudding,
you know to get their tupperware back to them
That's 68 words in 21.5 seconds, for a rate of 190 wpm, which is pretty much normal.
Here's the first clause:
oh it's wonderful at homecoming to go to the turnip green supper
In the course of setting this up as part of a gallery of prosodic patterns, I noticed some striking things about the segmental phonetics of that first phrase. I'll explain what it is tomorrow — meanwhile you may want to see what you come up with. There are several options.
Chips Mackinolty said,
April 5, 2017 @ 8:45 am
Curious to know what the evidence/source of 190 wpm as being close to normal:
"That's 68 words in 21.5 seconds, for a rate of 190 wpm, which is pretty much normal."
As a sometime speech writer I have always calculated, and counted on delivery, a much lower rate of 120-160 words a minute, but perhaps that is an artefact of delivering a "written speech" rather than unrehearsed/scripted/spontaneous speech. Would be really interested to know.
[(myl) See e.g. "Sex and speaking rate", 8/7/2006, which cites a table from Diana Binnenpoorte, Christophe Van Bael, Els den Os and Lou Boves, "Gender in Everyday Speech and Language: A Corpus-based Study", Interspeech 2005, which I describe this way: "…223 WPM for the males vs. 220 WPM for the females with pauses included, and 274 WPM for males vs. 266 WPM for females if pauses are excluded".
The same post also cites Jiahong Yuan, Mark Liberman and Chris Cieri, "Towards an Integrated Understanding of Speaking Rate in Conversation", ICSLP 2006, which gives these numbers for speaking rates in two English conversational corpora: "Call Home" 214 ± 6.73 wpm, and "Fisher" 193 ± 0.71 wpm. Those are overall rates, pauses included.
Or "The rhetoric of silence", 10/3/2004, where I note that in a presidential debate between G.W. Bush and John Kerry, "Bush's overall speech rate was slower (155 words per minute vs. 167 words per minute), but while the two men were actually talking, Bush talked considerably faster (220 words per minute vs. 202 words per minute)."
Or "How fast do people talk in court", 3/21/2009, which features this plot of a running (30-second) average during a sample deposition:
Or "Kennedy Speed: Fact or Factoid?", 9/15/2010, which presents this distribution of overall speaking rates in a large collection of conversational telephone speech:
]
Ellen K. said,
April 5, 2017 @ 9:14 am
The line in bold reads to me like a colloquialism, not a misnegation. And likely, as Mark Liberman notes, either an editing error or subtle humor. I assumed the latter when I read it, without considering the possibility of an editing error.
Y said,
April 5, 2017 @ 1:06 pm
The vowel of 'go' made me think for a moment I was listening to an Australian speaker. The long of supper was also unexpected. But then, I haven't had much exposure to Southern accents.
Y said,
April 5, 2017 @ 1:07 pm
Should be "The long <u> of 'supper'".
Rubrick said,
April 5, 2017 @ 4:31 pm
After that setup, I very much wanted to know what unstudied syntactic novelty you found in the first ten seconds of the clip. I can spot a few candidates, but I don't have the background to know what constitutes a novelty.
[(myl) "…the wager here is that if you examine a random stretch of English speech closely, within 10 seconds (of speech time, not your analysis time) you'll notice an instance of an interesting general pattern that has never been systematically studied in the linguistic literature."
An "interesting general pattern" is not necessarily syntactic. Hint: What I have in mind in this case has to do with phonology and phonetics.]
Daniel Barkalow said,
April 5, 2017 @ 6:57 pm
I'd imagine that what took until the lunch break wasn't observing something interesting, but checking with everyone ever to make sure they hadn't known it yet. I mean, that's not the sort of claim that I'd want to make without verifying.
Geoffrey K. Pullum said,
April 6, 2017 @ 1:05 am
It was of course a typo, and has already been corrected on the Lingua Franca post.
And of course lunchtime was the first opportunity to report on new discoveries to each other; discovery time (to the extent you can accurately locate a-ha moments or realizations of a generalization) could have been hours earlier.
Pflaumbaum said,
April 6, 2017 @ 4:23 am
Is it "go to" and "supper"? I hear something like Italian-style gemination of /t/ and /p/.
JPL said,
April 6, 2017 @ 4:57 am
"go" and "turnip" seem to have a low mid vowel [a] instead of a mid central vowel (or schwa), which is probably the more common variant for the first component of the diphthongized tense vowel or the "-er" stressed syllable nucleus. Later in "greens" in line 4 I could hear the same low vowel as the first component of the tense vowel; it's probably there in line 1 as well, and even the /o/ vowel in "homecoming" and the second syllable of "everybody" seem to have a similar vowel lowering relative to standard dialects.
JPL said,
April 6, 2017 @ 5:00 am
I meant "low central", not "low mid" in line 1 above.
Andrej Bjelaković said,
April 6, 2017 @ 6:43 am
Well I assume you're not referring to well-known Southern features, like the quality of GOAT (especially in 'go', not so much in 'home') or STRUT (in 'supper').
Maybe you mean what happens to different instances of /t/?
[(myl) You guessed it.]
Max said,
April 6, 2017 @ 10:57 am
The assimilation of <l w> to /w/ was something I hadn't noticed in American speech before.
Max said,
April 6, 2017 @ 10:58 am
(as in "school with")
Rubrick said,
April 6, 2017 @ 4:17 pm
@MYL: Ah, I see. The sentence prior to the quoted one ends with "…surprised that it took them until lunch time to come across something new in English syntax", so I assumed you were going to provide a syntactic, rather than phonetic, example yourself.
Bathrobe said,
April 7, 2017 @ 6:23 pm
In the transcript, there is a sentence that runs:
I'm sure
you'd know who mi- you'd know that Miss Layler's made the turnip greens
which doesn't make much sense. The reason for the strange transcription is, of course, English orthography.
To represent how it's heard by a speaker of English, it might be better transcribed as:
I'm sure
you'd know who may- you'd know that Miss Layler's made the turnip greens
But this is misleading in itself because it looks like the word 'may'. The alternative ma- would not be a very happy choice, either, because it looks like it's the first syllable (with so-called short 'a') of a word like 'matter'.