## I wouldn't be surprised if they didn't

David Craig sent in a link to yesterday's Blondie, a strip that I don't normally read. He may have uncovered the secret identity of Geoff Pullum's correspondent — but in any case, the last panel has a nice instance of overnegation, of a kind that I don't think we've discussed before:

Like several other species of apparent overnegation (e.g. "fail to miss", "cannot underestimate"), this one is so common that it appears to be in some sense a valid construction. It appears in BBC features as well as various local papers.

But the seeds have been sown and I wouldn't be surprised if she didn't try, once she's feeling more confident herself, to persuade him into the deep end.
"I wouldn't be surprised if we didn't see there are a couple of days with some good news and very, very positive market news," Houge said, noting potential days of 10 percent spikes.
"I wouldn't be surprised if you didn't lose $157,000 in taxes," Van Tuinen said. It was late, and I wouldn't be surprised if he didn't leave some of his audience back down the road somewhere impaled on point number 10 or 11. A quick search in Google Books verifies that it's common in popular novels: I wouldn't be surprised if she didn't try to make trouble, if she hasn't already, that is. I wouldn't be surprised if she didn't let herself get pregnant on purpose to hook Kent. Knowing how outspoken Ella was, I wouldn't be surprised if she didn't confront Casey about her identity. When that broke up, her husband conveniently kicked off — I wouldn't be surprised if she didn't put rat poison in his juleps — and she high-tailed it up here to grab up ol' Pom before somebody else did. The transcripts in The Kennedy Tapes (Ernest R. May & Philip Zelikow, 2002) show JFK using this construction (p. 219), in discussing reactions to the anticipated confrontation at the Cuban blockade line: Robert Kennedy: Well, there isn't any choice. I mean, you would have been, you would have been impeached. President Kennedy: That's what I think. I would have been impeached. I think they would have moved to impeach. I wouldn't be surprised if they didn't move to impeach right after this election, on the grounds that I said … and didn't do it … and let … I mean, I'd be … And Suzanne Marrs' biography of the great writer Eudora Welty quotes her saying about William Winter "I wouldn't be surprised if he didn't prove himself to be one of the best governors in the nation." On a literal interpretation, the intended meaning appears to be undermined semantically by the second negation (the one in the if-clause, which in this construction always (?) seems to be final). And the sentence seems to have the intended meaning if the second negation is left out — for example, "I wouldn't be surprised if you lost$157,000 in taxes", or "I wouldn't be surprised if they moved to impeach right after this election."  Still, this construction is clearly more than just an occasional slip.

It might be an example of negative concord in action, even if classical negative concord is a within-clause phenomenon (I think). But I wouldn't be surprised if some of you didn't have other interesting theories.

1. ### John Lawler said,

February 23, 2009 @ 8:28 am

It reminds me of the "spurious negation" I grew up with in DeKalb County, IL, which introduces a negative in so- affirmation tags that has absolutely no semantic content, like the one in (2) below, which is (I have to point out) synonymous with (1) in DeKalb County:

(1) Joe went to Chicago yesterday, and so did I. (normal so-tag)
(2) Joe went to Chicago yesterday, and so didn't I. (spurious neg so-tag)

Except this couldn't possibly be negative concord, since there's no negative to trigger it. The phenomenon has also been reported from Hawai'i and upstate NY (which is where DeKalb County was settled from originally).
More details available in "Ample Negatives", CLS 10, 1974.

2. ### AJD said,

February 23, 2009 @ 8:50 am

"So didn't I" is reported from Hawaii, DeKalb County IL, and Upstate NY? I had no idea—I've only ever heard (of) it from eastern Massachusetts.

Aviad Eilam has been doing research on "expletive negation" in Hebrew, and he argues that negation in a free relative clause carries the same semantics as "whatever" does in English: so the Hebrew sentence that appears to mean 'The journal printed what Joe didn't submit' actually means 'The journal printed whatever Joe submitted.'

3. ### James said,

February 23, 2009 @ 8:55 am

"So didn't I" is also common in Rhode Island. For a long time I thought it meant "Neither did I," which, as you might imagine, caused me some confusion.
I wonder if the RI and the IL usages have a common ancestor.

4. ### Mark Liberman said,

February 23, 2009 @ 9:03 am

AJD: Aviad Eilam has been doing research on "expletive negation" in Hebrew, and he argues that negation in a free relative clause carries the same semantics as "whatever" does in English…

"Expletive negation" is well documented in Romance languages, and is obviously relevant here. In fact, a bit of web search turns up a book chapter that seems to nail the whole thing: Wim van der Wurff, "On expletive negation with adversative predicates in the history of English", pp. 295-328 in Indrid Tieken-Boon van Ostade et al., Eds., Negation in the History of English, 1999.

He discusses examples from Late Middle English like

þan I haue no doute þat it ne schal wel kun telle þee of hem
then I have no doubt that it Neg will fully be-able tell you of them
'Then I have no doubt that it will be able to tell you all about them'

He explains that this construction "is an instance of what is usually called 'paratactic negation' (after Jespersen 1917: 75, 1924: 334) or "expletive negation" (a term that has its origins in French language studies …). It involves a verb or noun with the meaning 'fear', 'forbid', 'prohibit', 'hinder', 'prevent', 'avoid', 'deny', refuse', 'doubt' or another predicate with some kind of negative meaning, which triggers the use of a negative marker in a subordinate clause. What makes the construction interesting is the fact that the negative marker is semantically redundant, or expletive."

He adds that "Another interesting aspect of this construction, which should endear it to historical and comparative linguists, is its diachronic and cross-linguistic variability". As I said, this is clearly more than just an occasional slip.

5. ### Chris said,

February 23, 2009 @ 9:44 am

You're right that "I wouldn't be surprised if they both didn't end up…" is expletive negation: it's not "overnegation". The sentence in question is completely grammatical under the intended interpretation in standard English, unlike many of the examples of overnegation cited on LL, for which the intended interpretation is generally clear, but which are not grammatical under this interpretation in standard English. Moreover, and this is the interesting part, "I wouldn't be surprised if they both didn't end up…" is truth-conditionally equivalent to the equally grammatical "I wouldn't be surprised if they both ended up…".

Present-day standard English is actually pretty odd in allowing expletive negation in as restricted a range of contexts as it does: as far as I can tell, these are limited to 1) complement clauses of negated verbs of surprise, 2) the complement clause, introduced by if/whether, of 'let's see', e.g. 'Let's see if we can('t) make this point a bit clearer."

Can anyone think of any other contexts that allow expletive negation in PDE?

6. ### Mark Liberman said,

February 23, 2009 @ 9:55 am

Chris: "I wouldn't be surprised if they both didn't end up…" is truth-conditionally equivalent to the equally grammatical "I wouldn't be surprised if they both ended up…"

This is true of examples where the negation is "paratactic" or "expletive", but it's not true of such cases in general. Thus we can easily find examples where the second negative is independently meaningful, and even syntactically necessary, as in these examples from a Google Books search where it licenses a negative polarity item in the subordinate clause:

I wouldn't be surprised if you didn't come back at all.
Frank wouldn't be surprised if he didn't care for El Diablo either.

Similarly, there are cases where both positive and negative subordinate clauses are explicitly compared, to indicate the lack of any particular expectation:

I wouldn't be surprised if he did and I wouldn't be surprised if he didn't.

And "He wouldn't be surprised if she didn't even lock her doors" would mean something very different if we turned it into "He wouldn't be surprised if she even locked her doors".

So the negation in the subordinate clause is not always "paratactic", in sentences of this form.

Also, I'm not sure whether the construction under discussion is part of current standard written English. It's often criticized by language mavens — my web searches turned up four or five complaints, including ones by Lederer and Safire — and the examples that I found in major newspapers were all in quotations. It's clearly common in speech, including the speech of people like JFK and Eudora Welty, but it's likely to raise some eyebrows, at least, when written.

7. ### Rolig said,

February 23, 2009 @ 10:13 am

Am I right in thinking that expressions such as "I wouldn't be surprised if they didn't move to impeach…" (JFK), where the expectation is that the negated action will occur ("they will move to impeach") are semantically equivalent to expressions such as "I'd be surprised if they didn't move to impeach"? If so, could this have a parallel in the relationship between such expressions as

"I couldn't care less if they moved to impeach" and
"I could care less if they moved to impeach"?

8. ### Language Log takes a strange turn « Panther Red said,

February 23, 2009 @ 10:29 am

[…] for one welcome our new insect overlords this turn.  The last three days have seen  posts about Blondie, Doonesbury,  and Partially Clips.  Earlier this month, Grand Avenue and Get Fuzzy were […]

9. ### Dan Lufkin said,

February 23, 2009 @ 11:02 am

Anyone care to analyze "I couldn't fail to disagree with you less."? It's a favorite discussion-ender of mine. Dunno where I first heard/read it.

Dan

10. ### Robert Coren said,

February 23, 2009 @ 11:11 am

@Dan Lufkin: Of course constructions like this with the intent to amuse and/or confuse have also been around for a long time; I'm reminded of the story about some esteemed music teacher saying solemnly to the doting parents of a child who had just played for him: "The young lady is not without a certain lack of talent."

11. ### John Lawler said,

February 23, 2009 @ 11:31 am

And an awful lot of people who can count modulo 2 have bought into the double-negative rule, which is after all widely used productively without overnegation.

I must admit that I tend to interpret examples like the Blondie one as simple arithmetic mistakes, frequent enough in practice to be solidifying into a new set of constructions.

Oh, and New England areas are undoubtedly similar if not identical construction to "upstate NY". Probably there are other areas as well. Further deponent knoweth not.

12. ### marie-lucie said,

February 23, 2009 @ 11:33 am

In Kingsley Amis's Lucky Jim there is a pretentious character who speaks almost exclusively in such doubly negative expressions.

13. ### Dan Lufkin said,

February 23, 2009 @ 1:27 pm

Afrikaans has a kind of double negative where the negative (usually 'nie') is always doubled to make a marker for the beginning and the end of the negation. Example: Soms, maar nie in die afgelope uur nie. Sometimes, but not in the past hour not. (From a patient interview form I translated last week.) This construction can be very useful in a low-grammar language like Afrikaans.
This is attributed to the influence of Bantu (especially Khoi) on Afrikaans, but it's a construction I've also heard in dialect Swedish (with inte … inte). It's rare (in my experience) and is used for emphasis. (I know how inte moves around to show subordination, etc.). I wonder if it occurs (non-standard) in other Germanic languages.

14. ### Sili said,

February 23, 2009 @ 2:07 pm

It looks like blending of two constructions to my uninformed eye.

You suggest removing the last negation, "I wouldn't be surprised if they won", but it seems to me that "I'd be surprised if they didn't win" is equally wellformed.

So could this not have started with enough people simply changing their mind or forgetting which construction they wanted halfway through the sentence? Of course from there it has prolly spread to idiomatic usage.

15. ### acilius said,

February 23, 2009 @ 2:41 pm

I suspect Sili is exactly right. The example sentence in the post is certainly complex enough that its speaker might well have forgotten the precise wording with which she began well before reaching the end.

16. ### Andrew said,

February 23, 2009 @ 2:47 pm

Isn't the construction with 'let's see' a different kind of case? If we see whether something happens, we also see whether it doesn't happen; so those two constructions are equivalent in any case, without any special rule needing to be invoked.

17. ### Skullturf Q. Beavispants said,

February 23, 2009 @ 4:40 pm

Dan Lufkin at 1:27 pm: Does the Afrikaans example you cite simply mean something like "Sometimes, but not in the past hour"?

Superficially, this seems like something I've noticed among some native speakers of English (perhaps more common in the UK than North America): "Not in my house you don't." Meaning simply "Not in my house", where the repetition of a negative word is merely for emphasis and doesn't "deny the negative" and turn it into a positive.

18. ### marie-lucie said,

February 23, 2009 @ 4:51 pm

Skullturf,

It seems to me that even with a single clause intonation, there are two underlying clauses here, both incomplete:

(You do) not (do X) in my house,
you don't (do X in my house).

Does that make sense in more general terms?

19. ### Chris said,

February 23, 2009 @ 5:35 pm

Mark,

Of course you're right that "wouldn't be surprised" in a matrix clause doesn't *force* a negator in the complement clause to be interpreted expletively. I never intended to imply this. I was referring to the truth conditions of the example utterance in question. Right thinking folk nowadays don't think sentences in general (divorced from context) have truth conditions!

On the subject of the place of expletive negation in standard English, that's interesting that the language mavens have taken it upon themselves to excise this bit of grammar from our writing – I wasn't aware of this. I guess it's not surprising given that, to a woolly-minded prescriptivist unfamiliar with French or German say, expletive negation should seem the height of illogic. But I am surprised because, as we all know, prescriptivism tends to defend the grammatical system of the educated elite of the day (or perhaps a few decades past), however crazy its logic. And my subjective impression of expletive negation in British English is that it's very much a part of elite-speak and has no associations at all with lack of education or informality. Perhaps it's different in American English?

20. ### David J. Fried said,

February 23, 2009 @ 5:39 pm

Looking at the Espinal article cited by Mark Liberman, I find that one of the Spanish sentences discussed can be equally well expressed in English with overnegation:

to how many people not have+FUT.3sg killed this dictator
‘So many people must have been killed by this dictator!’

This could be more literally translated into English:

"How many people has this dictator not killed!" I think this is idiomatically preferable to "How many people has this dictator killed!"

I'm not linguist enough to advance the discussion properly. I will only say that this seems to me to be a rhetorical question, not an exclamation. And the force of the "not" is to elicit agreement from the hearer. In other words, it is equivalent to "This dictator has killed many people, has he not?"

Thoughts?

February 23, 2009 @ 6:43 pm

Thanks to Lance and AJD for bringing up my work in this context.

I'd suggest reading some of the literature; it addresses many of the issues posters have been discussing here. van der Wurff's (1999) paper on Middle English, which Mark mentions, is a nice place to start.

You can also begin with my paper, available at semantics archive:
http://semanticsarchive.net/Archive/mQ5ZjY3N/
I give a brief overview of the literature, and one of my first examples from Modern Hebrew is exactly like the one from the comic strip, but I distinguish it from the free relative and until-clause forms of expletive negation in Hebrew: the former is not grammaticized, while the latter is. I also deal with the rhetorical question vs. exclamative issue which David Fried notes.

Re the possible connection with negative concord: see van der Wouden (1997), who promotes precisely this idea.

22. ### John Swindle said,

February 23, 2009 @ 6:49 pm

Is the "spurious neg so-tag" of DeKalb County and elsewhere much different from this?

(3) Joe went to Chicago yesterday, and didn't I too!

23. ### Mark Liberman said,

February 23, 2009 @ 7:04 pm

Chris: …that's interesting that the language mavens have taken it upon themselves to excise this bit of grammar from our writing – I wasn't aware of this.

For example, Richard Lederer and Richard Dowis, Sleeping Dogs Don't Lay (2001), p. 12, under the heading "Double negatives are no-nos":

Another fairly common example of a double negative is "I wouldn't be surprised if she didn't show up unannounced" when the speaker obviously means "I wouldn't be surprised if she showed up unannounced".

Or William Safire, "Essay; Cuomo Vs. Clinton", NYT, 2/3/1992, where the speaker is the (in)famous Gennifer Flowers:

Forget the unrequited-love baloney from the woman giving "floozy" a bad name; in sound bites from Gennifer's Greatest Hits, she derogated Governor Cuomo, got a placating Governor Clinton to half-agree, and thereby breathed new life into the New Yorker's flirtation with the Democratic nomination.

"He seems like he could get real mean," she says about Mr. Cuomo on one tape snippet. A voice that is presumably Governor Clinton's murmurs "A mean son of a bitch . . ."

That could be construed as a compliment in political circles, requiring no apology, but then the entrapping caller observed: "Yeah, I wouldn't be surprised if he didn't have some Mafioso major connections." (Her double negative made her unintentionally accurate: Mr. Cuomo does not have Mafioso connections.)

Instead of dismissing it with "That rumor's been in all the newspapers for years," Governor Clinton said, "Well, he acts like one."

Mr. Safire is not being especially censorious about the construction here, but he clearly believes (as I do, at least on Tuesdays and Thursdays) that it means the opposite of what its speaker intends.

Similar remarks can be found in Michael Swan's Practical English Usage (1995), Lois Irene Hutchinson's Standard Handbook for Secretaries (1947), etc.

R.L. Trask's Say What You Mean, published posthumously in 2005, opts on p. 95 for the "poor monkey brains" hypothesis:

But even careful speakers sometimes get lost in syntactically complex constructions. Here is an example: *I wouldn't be surprised if the Government didn't call a snap election. What the writer means, almost certainly, is this: I wouldn't be surprised if the Government called a snap election. The second negative has wrongly crept into the original example, producing a result that appears to assert exactly the opposite of what was intended.

And Larry Trask was a first-rate linguist, whose death in 2004 at age 59 we mourned here. He might have been wrong in this case — at least I think he should have added a note about the historical origins of the construction — but he was no fool, to say the least.

Google Books also turns up Hubert Anthony Shands, Some Peculiarities of Speech in Mississippi, 1893, which includes the entry:

Wouldn't be surprised if it didn't. This expression is frequently used by all classes in the place of wouldn't be surprised if it did; as in speaking of the weather, one says: "I wouldn't be surprised if it didn't rain," meaning that he thinks it is going to rain.

And J.H. Combs, "Old, Early and Elizabethan English in the Southern Mountains", Dialect Notes (American Dialect Society, 1917), has this to say:

As among the Elizabethans, there is every variety of grammatical inaccuracy. The double negative has stoutly maintained its raison d'être among the Southern hillsmen, and we find a more extensive use of the negative with them than with the Elizabethans. Here is a bit of "juggling" with the negative: "I wouldn't be surprised if it didn't rain," (meaning, "I wouldn't be surprised if it should rain").1 "Rance, he ain't got nary none." "Fotch-on (educated) preachers ain't never a-goin' to do nothin' nohow. "I hain't never seen no men-folks of no kind do no washin' (of clothes)."

The laconic footnote reads "Certainly not local. –Ed." I'm not sure how to interpret this, exactly, but at a guess, the editor is pointing out that the footnoted expression is much more general in its usage than the others in the paragraph.

24. ### Skullturf Q. Beavispants said,

February 23, 2009 @ 7:54 pm

John Swindle at 6:49 pm: Thank you for the example!

I became aware of the "so don't I" phenomenon fairly recently, and have never encountered it in the wild. Intellectually, I understand that certain regions may have common expressions that sound "illogical" to outsiders but perhaps can't be called "wrong" if they are natural to those who have grown up with it and if the meaning is understood. However, subjectively, I had real trouble wrapping my mind around it: "You mean there really are people who say 'So don't I' for 'So do I'?!"

Your example reminds me that there are expressions more familiar to me that really aren't so different. Another one is "Don't I know it!" uttered with the intonation of a statement, not that of a question. Unlike "So don't I", I grew up familiar with "Don't I know it!", and effortlessly understood it to mean "I sure do know it!" without even really consciously noticing the "negativity".

25. ### Nathan Myers said,

February 23, 2009 @ 7:55 pm

I'm a big fan of the triple-negative. It challenges the double-negative nannies who must then bite their collective tongue when they ultimately discover the unaccountable correctness.

26. ### Mike Scanlon said,

February 23, 2009 @ 8:14 pm

But in the end, what she is saying is that it would not surprise her if they did not end up with academy awards; that is, as much of a performance as they went through it would not warrant an award, and so she would not be surprised if they didn't get it. She obviously would be surprised if they did, and so would I!

27. ### Robert Coren said,

February 23, 2009 @ 8:33 pm

I wonder if "Don't I know it!", although not intoned as a question, started life (however many centuries ago) as the corresponding rhetorical question.

I also wonder if the locution used by Eeyore during the Expotition to th North Pole fits in here somewhere:

"It will rain soon, you see if it doesn't," said Eeyore.

Roo looked to see if it didn't, and it didn't…

28. ### John Lawler said,

February 23, 2009 @ 8:59 pm

@Nathan:
Triple negatives are not without a certain je-ne-sais-quoi.

Whatnot is a member of a large and interesting class of words called "nonce terms" in the trade. Other nonce terms include doohickey, thingumajig, thingie, gadget, Whoozis, and What's'erface?. Some people use whatnot to mean the same as whatever in at least some contexts; but for most American English speakers a (single, individual) whatnot is a small home-decorating icon of at least questionable taste. A whatnot shelf is a small wall shelf for displaying individual whatnots.

@John Swindle:
No, the DeKalb County construction is not at all equivalent to (3). Sorry.
Full details (really) in the 1974 paper.

@Skullturf:
Yes, there really are such people. Like me; I happen to be a professional linguist who grew up speaking that particular micro-lect. It doesn't feel weird; it's just another natural way to talk, like "dropping your g's".

29. ### Nathan Myers said,

February 23, 2009 @ 10:35 pm

John Lawler: Thank you. I'm used to "whatnot" in the aggregate sense, recalling, perhaps, "cetera": . But how did it get its "-not" bit? My tingling folk-etymology-sense suggests it, like some of the examples cited above, evolved from (the implied answer to) a question. Merriam-Webster seems to confirm this with a cryptic "Etymology: what not? Date: 1540".

Perhaps M-W is confirming the suspicion voiced above that other "not" statements mentioned also originated as (rhetorical) questions.

30. ### Skullturf Q. Beavispants said,

February 23, 2009 @ 11:11 pm

Interesting. For me, "whatnot" has only the meaning similar to "whatever" or "what-have-you", and I was unaware of referring to an individual object as a "whatnot".

31. ### dr pepper said,

February 24, 2009 @ 12:34 am

I knew someone who said "so doesn't". She had spent her childhood in Nashua, NH. For what it's worth, she said the neighborhood was full of illegal french canadians.

32. ### acilius said,

February 24, 2009 @ 8:30 am

Mark Liberman: Thanks for mentioning that Larry Trask quote! That's what I was trying to remember when I wrote my response to Sili's comment #14 above, it's been bothering me ever since.

33. ### Ellen K. said,

February 24, 2009 @ 10:34 am

@ Mike Scanlon: I take that statement from Blondie in the comic as an exaggeration. She's praising his acting ability, not putting it down. Even though she doesn't, we presume, actually think they'll win academy awards, she exaggeratedly claims they deserve one. If she were trying to communicate, "well, they don't deserve an academy award" she wouldn't have worded it as she did.

And, as a more general response, I think, when this contruction is said aloud, tone of voice may make a difference. Not in how the "didn't" is said, but in the phrase as a whole. An exclamation, not a statement. A tone that indicates it expressed a feeling about what is being commented on, not a factual opinion.

34. ### bread & roses said,

February 24, 2009 @ 2:23 pm

I think the double negative in the strip is for tone, not a confusion halfway through the sentence. It makes the assertion a little less strong (as opposed to "not in my house, you don't!" which makes the assertion more strong). And I think it also makes it more purely rhetorical. In the larger context, Blondie *would* extraordinarily surprised if Dagwood won an academy award.

I think the doubling of the negative works, for me, like saying "why don't you see if you can find out the time" instead of "what time is it?"
(which, for me, has exactly the same meaning as "why don't you see if you can't find out the time" except that the added "can't" makes it even softer). It's more social, softer, more persuading and agreeable, bordering on wheedling. It reminds me of "might could" – "And then, we might could eat a little lunch". It adds more tentativity to indicate flexibility, agreeableness, openness to other possibilities and points of view.
The construction in the cartoon (as opposed to either form with only one negation) slightly changes the tone to emphasize the bond between the two women, for me; Blondie isn't saying something her friend doesn't know; she's expressing an opinion she believes is shared, not asserting something- just like "don't I know it!"

That's how it works in my head, anyhow.

35. ### blahedo said,

February 24, 2009 @ 3:57 pm

Interesting that a poster identified the overnegated "so don't I" as a DeKalb construction—I grew up in the Chicago area and came back to Illinois after grad school, and still only ever remember hearing it when I was out east, in RI and eastern Mass. However, at the time I conjectured that it was related to a more familiar construction "so don't we all", which sounds fine to me. I agree with Robert Coren that a derivation by way of rhetorical question seems likely (though the modern intonation as I've heard it carries no interrogative feel anymore).

36. ### Skullturf Q. Beavispants said,

February 24, 2009 @ 4:32 pm

bread & roses: Thanks for that other example, "Why don't we see if…" That's another expression that I have grown up with (unlike "So don't we all"), and I'm familiar enough with it that I almost don't "hear" the negative until someone points it out to me.

37. ### Will Shoemaker said,

October 22, 2009 @ 6:34 am

But what about the crazy verb tenses going on here? "I wouldn't be surprised if he went bowling tomorrow" — is that correct, or should it be "I wouldn't be surprised if he goes bowling tomorrow" …the latter sounds totally wrong…hrm..