Plural data
« previous post | next post »
Today's xkcd:
Mouseover title: "If you want to have more fun at the expense of language pedants, try developing an hypercorrection habit."
That should be "…developing another hypercorrection habit", since making data plural in that situation is exactly analogous to using whom in "Whom are you, anyways?". But then, as Ben Zimmer has pointed out to me, that would spoil the joke involved in the choice of an in "an hypercorrection".
Tom S. Fox said,
October 3, 2014 @ 5:05 pm
But then, as Ben Zimmer has pointed out to me, that would spoil the joke involved in the choice of an in "an hypercorrection".
Nice save, there.
Nathan said,
October 3, 2014 @ 5:16 pm
Data the android is singular, of course.
But data as in "information" is a mass noun.
Craig said,
October 3, 2014 @ 5:24 pm
He should have been called Lt. Datum.
The use of "an" before a word beginning with an aspirated H has always seemed odd to me, but I didn't grow up hearing it. I suppose it would make more sense when using the kind of British accent in which the initial H is silent, as in "an 'ypercorrection."
James said,
October 3, 2014 @ 6:01 pm
It's when the first syllable is unstressed that some people like to use 'an'.
She was an historian of renown, and she told me an hysterical joke.
For myself, I don't quite like either 'a' or 'an' for those words. It just doesn't sound good either way. But (I think) I say 'a'.
Gregory Kusnick said,
October 3, 2014 @ 7:04 pm
The thing is, though, that there apparently never was a time when an hysterical seriously dominated over a hysterical; they tracked pretty closely until about 1930, when a pulled into the lead.
Not so for an historian v. a historian, where an had a clear historical advantage, which it finally gave up around 1930.
So it seems there isn't (and wasn't) a general rule about unstressed initial syllables. Rather, many people make a special case for words derived from history that they do not make for unrelated H-words.
Jason said,
October 3, 2014 @ 7:14 pm
@Craig
The use of "an" before a word beginning with an aspirated H has always seemed odd to me, but I didn't grow up hearing it. I suppose it would make more sense when using the kind of British accent in which the initial H is silent, as in "an 'ypercorrection'.
It's always infuriated me, particularly with newsreaders at the Australian Broadcasting Corporation, who always make a point of saying, rather pretentiously, "an historic occasion", but never "an histogram" or "an healthy diet" or any other word starting with an "h" sound. What on earth would possess someone to conclude "an historic" is correct style but not apply it to /any other h-initial word in the language?/
Brett said,
October 3, 2014 @ 7:17 pm
I don't think the mouseover text is suggesting developing another hypercorrection habit in addition to the one shown in the cartoon. Rather, the cartoon represents a particular instance of what the mouseover text is suggesting.
Ralph said,
October 3, 2014 @ 7:33 pm
Brett, note "more fun" in the mouseover text. The mouseover suggests that the illustration shows one way to annoy pedants, and that hypercorrection is another different way. Mark is pointing out that Randall has not realized that data/data is itself an example of hypercorrection.
Ray Girvan said,
October 3, 2014 @ 7:41 pm
@Jason: What on earth would possess someone to conclude "an historic" is correct style but not apply it to /any other h-initial word in the language?
Etmyology is a factor. It has become particularly stuck to words with a French origin: history, hotel, horror …
Matt said,
October 3, 2014 @ 7:51 pm
He should have been called Lt. Datum.
Then we could have dialogue like "We'll wait for Datī report, and make our decision based on what we hear Datō. In the meantime, fetch Rikrem!"
Ralph said,
October 3, 2014 @ 8:00 pm
For my dialect (southern UK, never drop initial h) it appears that James nailed it in observing that "an" only sounds right when the first syllable is unstressed.
"an" is not acceptable for these in my dialect:
A History of Violence
a horror movie
a hypercorrective habit [major stress is on first syllable]
"an" works fine for these:
an historic occasion
an horrific movie
an hotel
an hypercorrection [major stress is on fourth syllable]
Justin said,
October 3, 2014 @ 9:23 pm
Is it really hypercorrection if you do as the comic says and make "data" plural only when referring to the character? I'd think it would only be hypercorrection if you made it plural in every instance; that to be hypercorrection, you have to be overcompensating in some way. The way I understood the term, using "whom" instead of "who" for objects alone is a correction, using it in every instance is a hypercorrection, and using it for subjects alone is just plain wrong.
Ralph said,
October 3, 2014 @ 9:43 pm
@Justin
Extending a correction algorithm to any inapplicable case is an instance of hypercorrection. I think it's a term for a type of error; you seem to be analyzing it as a psychological trait like hyperactivity.
SP said,
October 4, 2014 @ 12:10 am
But in "the most popular Star Trek character are Data," it's "character" that's being treated as plural, not "Data". Or am I just being the pedant that the joke is aimed at?
D.O. said,
October 4, 2014 @ 1:14 am
On a more abstract note, changing your speech to annoy pedants is also a form of hypercorrection.
Jamie said,
October 4, 2014 @ 2:54 am
@Craig
But to me, 'datum' is a completely different word with a different meaning (a fiducial mark) rather than the singular of information/data.
John Walden said,
October 4, 2014 @ 3:09 am
Like 'Ralph said' I'm a BrE speaker who doesn't drop his aitches but unlike him I never say or write an 'an' before an h that isn't silent (I do hope I've got those negatives sorted out!). In the case of 'hotel', those who treat it as a French borrowing seem perfectly justified in putting an 'an' in front of it. But before a well-formed 'h' it sounds, to me at any rate, like a rather Edwardian pronunciation of 'golf' as 'goff'. Or indeed of 'Edwardian' to rhyme with 'ed hard Ian'.
Pronounced in a more English way there is indeed a shift in stress from 'otEL to hOtel, for some people. But I don't see that stress rule working, especially not for 'an hypercorrection'. Which makes me wonder if there isn't a French connection: 'an historical novel' sounds merely slightly fogeyish while 'an hysterical novel' sounds plain strange.
GH said,
October 4, 2014 @ 5:45 am
@John Walden:
Nevertheless, that is (or was) the pattern. Fowler gives the examples "an historian", "an hotel", "an hysterical scene", "an hereditary title" and "an habitual offender" (though even in his day he considered it old-fashioned and anomalous, destined to disappear). "An historical" probably only sounds less stilted than "an hysterical" to you because you've heard it more often.
Personally (as a second-language speaker, but I must have picked it up in England), I believe I would only use it when the word has three or more syllables, so not for "hotel" (the only bisyllabic example I can think of… hmm, "Hanoi"?). This article argues, plausibly, that this is because it is more natural when the first syllable has only tertiary stress.
GeorgeW said,
October 4, 2014 @ 7:24 am
@GH: hoagie, hobbit, hobby, hobo, hockey, holdout, hogwash, . . . . .
GH said,
October 4, 2014 @ 7:32 am
All of those have stress on the first syllable, GeorgeW.
GH said,
October 4, 2014 @ 7:38 am
Just noticed the comment system ate my link. This is the article that discusses the difference in first-syllable stress between "hotel" and "historical": http://grammartips.homestead.com/historical.html
GeorgeW said,
October 4, 2014 @ 9:34 am
@GH: Sorry, I missed that restriction.
Rodger C said,
October 4, 2014 @ 11:04 am
My mother (b. 1920, Eastern KY) used to say "an hibachi."
Jerry Friedman said,
October 4, 2014 @ 11:27 am
One rule that some people use for words beginning with an "h" (according to discussions at alt.usage.english) is to use "an" if the first syllable is unaccented and contains a short vowel. Thus "an hibachi" but "a hydraulic press". Another is that words from Greek and maybe Spanish as well as words from French get "an". Then some people say that English has three pronunciations of the letter "h": strong as in "hope", weak as in "historian", and silent as in "heir", and they use "an" before the weak and silent ones.
I think that a lot of people who use "an", especially Americans, feel the older and more British form is more appropriate for solemn topics such as heroic deeds, or to show possibly genuine respect as for Hispanic people, that for ordinary topics such as hilarious jokes and Hawaiian shirts (see Google ngram result). That sounds terrible to me.
Rebecca said,
October 4, 2014 @ 7:02 pm
Re Lt. Data.
No reason to think an individual can't have a plural or mass noun name. I've had a mouse named Salt and cat named Toes. And then there's Pickles Sorrel.
Justin said,
October 4, 2014 @ 9:49 pm
@Rebecca:
But you never said "I own mice named Salt and cats named Toes", did you? That is, just because the pets were named with a plural word, you didn't treat the name as a grammatical plural as in the comic, did you?
John Walden said,
October 5, 2014 @ 2:59 am
The choice of "a piece of data" or "item of data" is consistent with "data" as an uncount singular, as far as I can see. It works like "furniture", news" or "gossip".
However, the combo of "The data are" and "One piece of data is" seems odd. You're stuck with datum. Are there examples of "a piece of" followed by a plural?
To some extent saying "The data are unreliable" nowadays is a display of erudition, to show that the speaker/writer knows things that hoi polloi are unaware of.
Paul said,
October 5, 2014 @ 3:59 am
BrE, never say "an hotel", fairly sure I would say "an hysterical joke" and almost, not quite, drop the h. And I'm not an (an!) habitual dropper of aitches. Interesting.
Rebecca said,
October 5, 2014 @ 11:16 am
@Justin,
No, of course not. I was (unclearly) referencing the comment above that the character "should" be Lt. Datum. While "data" might be a target of peeving about correct plural use in general, it's use as a character name is not a very good spot for this because that's a spot where plurals can function as grammatically singular names. In other words, you can't peeve that people don't get that "data" is plural, because that's a spot where any old plural-used-as-singular-name is perfectly ok .
Keith M Ellis said,
October 5, 2014 @ 2:10 pm
"What on earth would possess someone to conclude 'an historic' is correct style but not apply it to /any other h-initial word in the language?/"
Why on earth would you insist on such a categorical consistency?
Toma said,
October 6, 2014 @ 10:32 am
I remember a MASH episode when Col. Blake was introducing Pierce and McIntyre to someone:
Col. Blake: "These are Captains Pierce…"
Pierce (interrupting): "…and these are Captains McIntyre."
BZ said,
October 6, 2014 @ 12:54 pm
"An historic" (both print and spoken) is very common in US news sources as well. I can almost say I've heard "an historical" and "an hysterical" too, but I can't be sure. I don't think normal people around me talk like that, though. Also, I'm pretty sure these are the only examples of this phenomenon. They don't generalize to any other word regardless of what sound comes after the aspirated "h" and whether it's stressed.
Dan said,
October 6, 2014 @ 1:08 pm
The doctor on the original series were Bones, right?
GH said,
October 6, 2014 @ 1:55 pm
How about "an heroic" (effort, etc.)? There's currently one hit on Google News from the LA Times, and Google Ngram Viewer shows it at about 10% of total instances.
GH said,
October 6, 2014 @ 1:56 pm
… in AmE, that is.
BZ said,
October 6, 2014 @ 2:24 pm
@GH,
No, doesn't ring a bell
J. W. Brewer said,
October 6, 2014 @ 6:45 pm
Maybe this is taking a joke-within-a-joke too seriously, but is using an archaism (even incongruously because in a phrase involving a recently-coined lexical item) really a type of hypercorrection, or is it something else (with "hypercorrection" requiring something that would have been Just Wrong in all conceivable time periods, regional varieties etc.)? Because if you go back far enough, "an" with h-initial words is certainly found even with first-syllable stress, as in e.g. "And an highway shall be there" (part of Isaiah 35:8 in KJV). The U.S. Constitution, which is not nearly as old as the KJV, uses the bigram "an uniform" which sounds even weirder to my ear than "an" with anything h-initial, but was apparently (per a quick consultation with the n-gram viewer) the majority variant until the 1820's.