Q: I often encounter a sentence such as “I wouldn’t be surprised if she didn’t steal the necklace,” when it actually means the opposite—the speaker or writer wouldn’t be surprised if she DID steal it. Is there a term for this (a type of double negative, maybe)? And how did it come to be so widespread?
A: We’ve seen several expressions for this kind of construction. Terms used by linguists today include “expletive negation,” in which “expletive” means redundant; “negative concord,” for multiple negatives intended to express a single negative meaning; and, more simply, “overnegation.”
Yes, it’s also been called a “double negative,” the term H. L. Mencken used for it more than 80 years ago. Like linguists today, Mencken didn’t find this particular specimen odd or unusual. As he wrote in The American Language (4th ed., 1936), “ ‘I wouldn’t be surprised if it didn’t rain’ is almost Standard American.”
The linguist Mark Liberman discussed this usage—“wouldn’t be surprised” followed by another negative—on the Language Log in 2009. He called it a “species of apparent overnegation” along the lines of “fail to miss” and “cannot underestimate.” (More on those two later.)
Of course, what appears to be an overnegation may not be so. For instance, if everyone but you is predicting rain, you might very well respond with “I wouldn’t be surprised if it didn’t rain” (i.e., you wouldn’t be surprised if it failed to rain). No overnegation there, just two negatives used literally, nothing redundant.
But the usage we’re discussing is a genuine redundancy with no literal intent. And it’s a type of redundancy that’s very common, especially in spoken English. Yet it seldom causes confusion. People generally interpret those dueling negative clauses just as the writer or speaker intends.
You’re a good example of this. While you noticed the redundancy there (“I wouldn’t be surprised if she didn’t steal the necklace”), you correctly interpreted the writer’s meaning (if she did steal it). And no doubt most people would interpret it that way, whether they encountered the sentence in writing or in speech. Why is this?
In the case of written English, our guess is that readers interpreting the writer’s intent take their cues not only from the surrounding context but also from their own past experience. They’re used to seeing this construction and don’t automatically interpret it literally.
In the case of spoken English, where the usage is more common, listeners have the added advantage of vocal cues. Take these two sentences, which are identical except for the different underlined stresses. A listener would interpret them as having opposite meanings:
(1) “I wouldn’t be surprised if he didn’t win” = I wouldn’t be surprised if he won.
(2) “I wouldn’t be surprised if he didn’t win” = I wouldn’t be surprised if he lost.
In #1, the redundant or overnegated example, the speaker emphasizes the verb and whizzes past the superfluous second negative (“didn’t”). But in #2, the literal example, the speaker emphasizes the second negative, so there’s no doubt that it’s intentional and not redundant.
Language types have been commenting on the overnegated “wouldn’t be surprised” usage since the 19th century.
On the Language Log, Liberman cites this entry from “Some Peculiarities of Speech in Mississippi,” a dissertation written by Hubert Anthony Shands in 1891 and published in 1893: “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.”
The usage wasn’t peculiar to Mississippi, though. In old newspaper databases, we’ve found 19th-century examples from other parts of the country.
These two 1859 sightings, the earliest we’ve seen, appeared in a humorous story, written in dialect, from the May 7, 1859, issue of the Columbia Spy, a Pennsylvania newspaper:
“ ‘There’s been so much hard swearin’ on that book’ (pointing to Logan’s Bible) ‘I wouldn’t be surprised if the truth was not pretty considerably ranshacked outen it.’ ”
“ ‘I wouldn’t be surprised if you wa’nt vain arter this.’ ”
This example is from newspaper serving the twin cities of Bristol in Virginia and Tennessee: “I wouldn’t be surprised if some of them didn’t run away after all without paying their bills.” (The Bristol News, Feb. 8, 1876.)
And here’s one from the Midwest: “The business interests of Salina feel the weight of their power, and we wouldn’t be surprised if even Nature did not pause for a moment and measure their colossal proportions.” (The Saline County Journal in Salina, Kansas, Jan. 25, 1877.)
As mentioned above, there are other varieties of overnegation besides the “wouldn’t be surprised” variety. Here are some of the more common ones, along with their intended interpretations.
“You can’t fail to miss it” = You can’t miss it
“We can’t underestimate” = We can’t overestimate
“Nothing is too trivial to ignore” = Nothing is too trivial to consider
“I don’t deny that she doesn’t have some good qualities” = I don’t deny that she does have some good qualities
“We don’t doubt that it’s not dangerous” = We don’t doubt that it is dangerous
As we’ve said, even readers or listeners who notice the excess negativity will understand the intended meaning.
The Dutch linguist Wim van der Wurff uses the term “expletive negation” for usages of this kind. As he explains, the first clause “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.” What follows is a subordinate clause with “a negative marker” that’s “semantically redundant, or expletive.”
He gives an example from a letter written by Charles Darwin: “It never occurred to me to doubt that your work would not advance our common object in the highest degree.” (From Negation in the History of English, edited by Ingrid Tieken-Boon Van Ostade and others.)
Historical linguists have shown that this sort of overnegation exists in a great many languages and in fact was a common usage in Old English and early Middle English.
“Negative concord has been a native grammatical construction since the time of Alfred, at least,” Daniel W. Noland writes, referring to the 9th-century Saxon king (“A Diachronic Survey of English Negative Concord,” American Speech, summer 1991).
But after the Middle Ages, the use of overnegation in English began to fall off, at least in the writings that have been handed down. Little by little, from around the late 15th to the 18th century, multiple negations became less frequent until they finally came to be considered unacceptable. Why?
Don’t point to the grammarians. It seems that this transition happened naturally, not because people started to object on logical or grammatical grounds.
In her monograph A History of English Negation (2004), the Italian linguist Gabriella Mazzon says the claim “that multiple negation was excluded from the standard as a consequence of the grammarians’ attacks is not correct, since the phenomenon had been on its way out of this variety [i.e., standard English] for some time already.”
As for today, Noland says in his American Speech paper, this type of overnegation “still has a marginal status even in standard English.”
We wouldn’t be surprised!