Irreproducible results
« previous post | next post »
This morning's Non Sequitur is not, as far as I know, a comment on my previous post:
Nor is this morning's Dilbert:
August 11, 2010 @ 8:48 am · Filed by Mark Liberman under Linguistics in the comics
« previous post | next post »
This morning's Non Sequitur is not, as far as I know, a comment on my previous post:
Nor is this morning's Dilbert:
August 11, 2010 @ 8:48 am · Filed by Mark Liberman under Linguistics in the comics
Powered By WordPress
groki said,
August 11, 2010 @ 12:08 pm
I didn't even know data can be real.
past "didn't" with present "can" sounds a little funny to my ears: technically correct but quite close to the boundary with incorrect. I would have used "could" myself. compare "*I didn't think you care" to "…you would care" or "…you cared."
maybe pointy-haired boss is supposed to sound less competent?
[(myl) It sounds fine to me — "can" rather than "could" makes it more generic, I think. Here are some real-world examples from web search:
Hobbs said the clash occurred because the rookie unaccustomed to press coverage didn't know he can be hit within 5 yards of the line of scrimmage.
I didn't know which daemons can be backgrounded and which not.
I didn't know a caesarean can be very dangerous for the baby.
I'm going to try with your blocking scheme: I didn't know blocks can be internal.
I didn't know kangaroos can be purple.
I bet you didn't know pillbugs can be trained to do that…
I didn't know tarragon can be brewed as tea.
FOr me, switching "can" to "could" modifies the meaning slightly — for the worse, in these cases and in the cartoon as well.]
Michael Straight said,
August 11, 2010 @ 12:25 pm
"I didn't even know data could be real" sounds to me like "I've seen lots of fake data, I didn't know there could be real data too."
"I didn't even know data can be real" sounds more like "What? Some of that data that flew by was real? Who knew?"
J. W. Brewer said,
August 11, 2010 @ 12:45 pm
I find Pointy-Haired Boss Dude's use of "can" rather than "could" highly unidiomatic-sounding. Maybe even ungrammatical in Standard American English? It seems parallel to a guy proclaiming his love by saying "I didn't know a woman can make me feel like this."
[(myl) Some examples from Google Books where "sequence of tenses" is similarly ignored:
He didn't know how vindictive you can be.
I didn't know how serious Parkinson's can be.
I didn't know family curses can be broken because of the Cross of Yeshua!
I'll bet you didn't know permalinks can be pretty, did you?
"Didn't you know crows can be harbingers of bad news?"
Didn't you know any citizen's pocket phone can be utilized as a bug?
I just didn't know that you can be more lonely in London than in the sleepiest country village.
I didn't know that promiscuity can be a symptom of clinical depression.
Do these also seem unidiomatic and even ungrammatical to you? I'm not questioning your right to feel that way, I'm just curious about what the facts are, and especially about whether there might be a "mind set" problem here.]
J. W. Brewer said,
August 11, 2010 @ 12:48 pm
I mean, presumably the joke is that he had not previously been genuinely uncertain and openminded on the question of whether or not data could be real, he had to the contrary been operating on the assumption that it couldn't be.
Mr Punch said,
August 11, 2010 @ 1:32 pm
My reaction: Idiomatic but slightly ungrammatical; tenses should agree. "Could" is proper as the past tense of "can," not as a conditional.
Mary Bull said,
August 11, 2010 @ 2:11 pm
I agree with myl that "could" substituted for "can" in the construction "I didn't even know data can be real" gives the statement a slightly different meaning — and one that doesn't fit the cartoon so aptly. Nor would it be a good change for the other examples he shows.
What J.W. Brewer says, though, is interesting. (It seems parallel to a guy proclaiming his love by saying "I didn't know a woman can make me feel like this.") I agree that "can" feels a bit "off" to me in his example sentence. But, "I didn't know that women can make a person feel like this" works for me.
So, what's going on with me, I wonder? (I'm a native speaker of American English.) Is it that I need the subject of the dependent clause to be plural, to match up present "can" to past "did"? I tried that, in J.W. Brewer's sentence, and it still wasn't comfortable. Changing "me" to "a person" does the trick. But only for the plural subject, not when I keep "a woman" but change "me" to "a person."
I do know that "rule" about matching up tenses, but I've got a feeling it has a lot of exceptions.
language hat said,
August 11, 2010 @ 2:41 pm
I find Pointy-Haired Boss Dude's use of "can" rather than "could" highly unidiomatic-sounding.
Groki and I agree with you, but apparently others find them equally acceptable though with slightly different meanings. Thus does language change.
Michael Straight said,
August 11, 2010 @ 3:33 pm
I think maybe using "can" this way is a shortened form of something like this:
I didn't know that the statement "data can be real" was true.
I didn't know that there was a rule: "you can be hit within 5 yards of the line of scrimmage."
I bet you didn't know this fact: "pillbugs can be trained to do that."
Haven't you heard the saying, "Crows can be harbingers of bad news?"
Will said,
August 11, 2010 @ 4:32 pm
I'm with groki, J. W. Brewer, and language hat. That sentence definitely gave me pause while I was reading it, and registered a "this sounds wrong". To me, it feels like it's on the verge of ungrammatical.
Circeus said,
August 11, 2010 @ 4:38 pm
Maybe it's the fact this is a very short statement with little adjoining text that make the use of "can" so salient? If it were embedded deep into a more complex text, we might be far more liable to let it slide.
Michael Johnson said,
August 11, 2010 @ 4:44 pm
The line sounds fine to me. Isn't this a standard case where sequence of tense is violated because the state described in the content clause persists to the time of utterance? As in:
John told [past] me that Mary is [present] pregnant.
Clayton Burns said,
August 11, 2010 @ 5:24 pm
[Deleted due to irrelevance]
J. W. Brewer said,
August 11, 2010 @ 6:52 pm
Re myl's examples, some of them feel weirder to me than others. I'm not sure if (w/o more time than I have right now) I can discern or articulate a pattern of my reactions to which of those sentences are where on the feeling-weird continuum, but I'm obviously not holding strictly to a sequence-of-tenses thing. Maybe it's more of a degrees-of-conditionality thing, or, as I tried to say in an earlier post, more a sense of surprise. I.e., the "can" seems more appropriate when it's just new information, as opposed to new information that contradicts a prior deeply-held belief (in this case, by hypothesis, the inherent bogosity of all so-called "data"). On the other hand there's certainly some syntactic thing going on, because I would have no problem with Pointy-Head saying "Wait a minute. Data can actually be real? I had no idea!"
I also agree that Mary Bull's recasting of my example sentence to make it seem less odd-sounding works (i.e. I find the "can" in her rewrite less problematic). Eternal fame and glory to whoever can articulate a theory for why that should be so. It perhaps may have something to do with how personally/directly invested the speaker was in the prior belief which has now been changed by new information? Indeed, I find the "can" examples from myl more acceptable if I stick in emotionally-distancing quotation marks, making them sound more like "I had hitherto been unaware of the truth of the zoological proposition: 'kangaroos can be purple.'"
Rubrick said,
August 11, 2010 @ 7:25 pm
I think Circeus has something there; it sounds worse to me with "even" removed: "I didn't know data can be real".
"I didn't know Language Log discussions can be so interesting" sounds very wrong to my ear.
Richard said,
August 11, 2010 @ 10:52 pm
Interestingly, "didn't know . . can" flew by me the first time, but I would always _write_ "didn't know . . could".
Jerry Friedman said,
August 12, 2010 @ 12:50 am
@language hat: Which way is the language changing, toward accepting can or forbidding it?
@Michael Straight: Is the difference between your two paraphrases that the "Who knew?" one indicates more surprise?
(After twenty years in a roly-poly region, I'm pleased to encounter a fellow pillbug speaker.)
@myl: In the original sentence, I could only use could. In most of your examples, can seems possible.
As far as I can tell, in a question beginning Didn't you know I'd always prefer can.
Here's one of your examples where I see a difference in meaning.
I just didn't know that you can be more lonely in London than in the sleepiest country village.
With could, it strikes me as literal: the speaker didn't know it was possible. With can, it happened to the speaker. (I believe. Are syntacticians immune to this phenomenon where the more you think about a locution, the less intuition you have for it?) This might be related to something else Michael Straight said. With can, it's almost a mention rather than a use. I get a feeling of "I didn't know I'd have to apply this statement to myself: 'You can be more lonely in London than in the sleepiest country village.'"
Speaking of sleepy…
Bob Ladd said,
August 12, 2010 @ 2:49 am
I'm with Richard – I didn't notice can the first time I read it, but agree it sounds pretty weird once I think about it.
I doubt this is a matter of language change, but rather a borderline case in a stable grammar. The whole issue is extensively discussed in Huddleston & Pullum's Cambridge Grammar on pp. 155-158, including a page and a half on "Factors affecting the choice between backshifted [e.g. could] and non-backshifted [e.g. can] versions".
Kapitano said,
August 12, 2010 @ 6:15 am
The pointy-haired boss is perfectly grammatical – but fairly uncommon. There's no reason for tenses to match between main and subordinate clauses, but educated speakers tend to make them match.
In my ESL work I've worked with many otherwise competent teachers who insisted that grammatically impeccable forms they didn't personally use were ungrammatical. Actual examples:
"I don't have some money."
"You mayn't do that here."
"We used not to go."
"You shouldn't've done that." (in writing)
There's nothing wrong with the grammar; They're just forms most native speakers usually avoid.
SeanH said,
August 12, 2010 @ 8:13 am
"We used not to go."
I never know what to do with this. "used not to" is the first thing to occur, but it feels 'incorrect'. "We usen't to go" feels 'correct' but awkward. I tend to just rephrase around it.
Gary said,
August 12, 2010 @ 8:24 am
@SeanH: I write "We didn't use to go" and hope for the best—>that no misguided editor insists it should be "We didn't used to go".
Diane said,
August 12, 2010 @ 8:59 am
I would definitely say "We used to not go."
Ken Brown said,
August 12, 2010 @ 8:59 am
I think I probably say "we didn't use to go" rather than "we didn't used to go". But how can I be sure? I have no memory of writing it and they sound exactly the same!
OK, "use to" and "used to" might be distinguished in careful speech but I'm pretty sure they they overlap a lot in normal speech. Either might be something like ['juːzdə] or ['juːstə] – and I'm not quite sure which I'd really say. Introspection isn't a very good way of finding our how you speak, is it?
Will said,
August 12, 2010 @ 1:57 pm
@Gary, I'm not an editor or a writer, but I definitely write "didn't used to" as opposed to "didn't use to", but I don't think the latter form is an error. And based on a quick google search, both forms are common and well-attested — I guess they are just equivalently valid alternate spellings. And I certainly wouldn't notice a difference when reading unless I was looking for it.
@Kapitano, the first three of those forms listed aren't just uncommon — they really are very awkward, despite their grammaticalness. I wouldn't fault an ESL teacher too much for insisting that they are ungrammatical. The fourth example ("shouldn't've") I agree is just uncommon (but grammatical), and not awkward, but at the same time I don't think it's valid, on spelling grounds ("shouldn't've" is not a standardly accepted spelling to my knowledge — the standard way to spell the expression pronounced that way is "shouldn't have").
Michael Straight said,
August 12, 2010 @ 3:49 pm
@Jerry Friedman, sorry my paraphrases did a poor job of expressing how I heard the two phrases.
To me, the "can" version makes the boss seem more removed. Like "data" is this thing he's heard rumors about and one of the rumors he'd never heard before is "data can be real."
Whereas "could" makes it sound like the boss knows data, has used data, it's always been fake in his experience, and he's surprised to learn that data could be real.
Chris said,
August 12, 2010 @ 4:46 pm
To my ear, the difference between sentences where "can" sounds OK and those where it doesn't depends on whether the part of the sentence being discussed can stand as a correct and independent sentence by itself.
Thus, to me, "I didn't know a caesarean can be very dangerous for the baby," sounds fine because "A caesarean can be very dangerous for the baby" can stand alone.
And to me, "I didn't know which daemons can be backgrounded and which not," does NOT sound right, because "Which daemons can be backgrounded and which not" can't stand as a complete sentence.
(There must be technical terms for this but as a lurker I will confess that any formal study of grammar is decades behind me.)
Xmun said,
August 12, 2010 @ 4:55 pm
@Ken Brown:
I'm pretty sure that I read in S. Greenbaum's Oxford English Grammar that he considers "didn't used to" as "perhaps preferable" (or some such expression) to "didn't use to". The judgment seemed to me so wide of the mark that it's stuck in my head.
The Cambridge Grammar (p. 115) more reasonably notes that "the spelling _used_ is sometimes found" in negative and interrogative constructions.
Justin L said,
August 12, 2010 @ 6:56 pm
I'm not sure of the proper terminology, but could it have something to do with subjunctivity/modality? Compare:
?I knew the sky is blue
?I knew the sky will be blue
?I knew the sky may be blue
I knew the sky must be blue
I knew the sky can be blue
To me all of these constructions express the condition that something is theoretically possible, but has not been experienced. When put in the past tense (i.e. "was blue"), it indicates that the condition has been experienced.
Steve Harris said,
August 13, 2010 @ 5:15 am
"I don't have some money."
Just wrong—ungrammatical in my idiolect. It has to be "I don't have any money" (if the intent is to say "I have zero money") or, e.g., "I don't have a dollar" (if the intent is to say "I have < $1").
"You mayn't do that here."
Very British to my ears—completely unheard of in AmE. I wouldn't call it ungrammatical, so much as incomprehensible in speech (though perfectly plain in writing).
"We used not to go."
Ungrammatical in my idiolect, though conceivably okay in BrE. "We didn't use(d) to go" or "We used to not go" are the only possibilities I would admit. (I can't reliably choose between "use" and "used" in the first alternative; I think it has to be "use", but there is very strong temptation to write it as "used".)
"You shouldn't've done that."
Looks very awkward in writing. I can see no reason for using a double-contraction in writing (unless, of course, it's direct dialogue). I can't say it's wrong, but I'd surely recommend against it as unidiomatic for writing—and to be idiomatic is the only reason for contractions in writing, so that's a Very Good reason not to use it.
tablogloid said,
August 13, 2010 @ 12:10 pm
Stay tuned for the premiere of prime time television's unreality show,
"So You Think You Can Data" .
Julie said,
August 14, 2010 @ 3:17 am
To me that opening sentence jumped out as Wrong. And until I read this column, I didn't know there were differing opinions on this. To use the present, I would have to use: "Data can be real? I didn't know that!"
"I don't have some money." It's certainly not idiomatic. Normally that sentence requires "any." However, we sometimes repeat back phrases from another sentence for emphasis, so there are occasions when a native speaker might use that sentence.
"You mayn't do that here." Not my dialect, but I wouldn't blink at it if spoken in a relatively unfamiliar accent. Somebody probably says it, somewhere.
"We used not to go." Odd, and I can't quite place the meaning. It seems like we used to refrain from going, as an intentional act? I'd say "We didn't use to go," for the common meaning. In writing, I might consider "we never (or hardly ever) went."
"You shouldn't've done that." That sentence sounds fine to me, and perfectly idiomatic. Certainly not ungrammatical, although the spelling indicates an informal context. My fictional characters might say it just like that. It's easier to read than "shouldn'a" which is what it sometimes sounds like.
tablogloid said,
August 15, 2010 @ 12:52 pm
@ Julie: A lofty English quote from Anthony Powell's "Dance to the Music of Time":
"Used you to see him often?"
un malpaso said,
August 16, 2010 @ 12:30 am
I think "can" sounds clunky here, but I also think Adams just chose it because it fit better. The design of a comic panel often forces artists into making awkward linguistic choices for the sake of composition.
Peter LeFanu Lumsdaine said,
August 16, 2010 @ 5:14 pm
As a speaker of fairly "high" BrE:
"We used not to go", or "…usedn't…", sounds and looks natural, a little formal; "we didn't used to go", also natural, more informal. "We didn't use to go" seems very unnatural, although it's certainly more logical, and google gives approximately equal hits (19m vs 22m). To my ear, "used" as an auxiliary has become distanced from other senses of "use" to the point of being essentially indeclinable.
I'm somewhat surprised there hasn't been a full LL post on this issue yet, thought it's cropped up several times in comments! I'd be interested to know how it varies across dialects, and historically. The OED's citations ("use", mainly sense 21 but the preceding senses seem relevant) paint a large and complicated picture, raising more questions than they answer…
Pflaumbaum said,
September 25, 2010 @ 10:03 am
Hmm. I'm also a speaker of 'fairly high' BrE (maybe not as 'high' as you, Peter… middle-class North London, educated to graduate level, BBC-accented with the odd inherited northernism). But to me
(1) We used not to talk about it.
(2) We used to not talk about it.
(3) We didn't use to talk about it.
(4) We didn't used to talk about it.
all sound natural, though they might have slightly differences nuances. I'm not sure there's any difference between (3) and (4) in speech, unless spoken very emphatically
Goes to show that we shouldn't take people's statements of 'feel' – including our own – to apply to speakers generally (or even fairly specifically!).
As for 'you mayn't', which to Steve Harris sounds 'very British', it sounds formal to the point of archaic to me, and I'm not sure I've ever heard it. If I have, I think it could only have been from someone very posh.