Prophylactic over-negation

« previous post | next post »

Almost the end of January, and not a single Language Log reader hasn't failed to complain about the lack of over-negation in any of this year's posts. But here's some naughtily nutty negation anyway:

"It's not that I don't doubt the sincerity of their desire to protect the talent. And believe it or not, we have the same ambition," Christian Mann, general manager of Evil Angel Productions who also serves on the porn industry's Free Speech Coalition, said last week after the council's vote. "We just don't believe their way is the best way." (Associated PressLA mayor signs law requiring condoms in porn films, Jan. 24, 2012; widely syndicated story.)

Hmm. That's a curious lack of non-self-doubt. So does it mean Mann does in fact doubt the sincerity  of "their" desire to protect the talent? I don't think so.

It turns out that the pattern is surprisingly common. Here are a bunch of web examples (from among many more, easily found) all of which appear to me to have a tad more negation than their intended truth conditions warrant:

It's not that I don't doubt Kevin is low enough to get Britney pregnant for some extra cash, it's just that I don't think he's clever enough.

It’s not that I don’t doubt that little happy crazy girl who’s smile was so big and bounced into a room is gone, it’s just completely dulled with that girl who smiles awkwardly when she’s forced to and walks with feet shuffling. It’s so bad that when I’m with her, even I doubt her capacity for depth of emotion.

It’s not that I don’t doubt color affects us profoundly.  We make choices based on colour on a daily basis – what clothes will we wear? What food will we eat?

It's not that I don't doubt colloidal silver's power, as it was used for centuries to treat wounds.

It’s not that I don’t doubt some of the books are good (or impacted literature or culture at the time); it’s just that there tends to be too much hype surrounding them. War and Peace? Couldn’t even finish. Jane Austen? Not too impressed. That being said, I did enjoy A Handmaid’s Tale (I have a terrible soft spot for dystopia).

You might say that Christian Mann and all these other guys are following a principle of prophylactic over-negation:

To protect your interlocutor from something false, swaddle the offending proposition in as many layers of negativity as possible.

So the idea is that speakers who produce all these negations might be subconsciously operating on the intuition the more layers of protective negativity you wrap around a proposition, the less chance that any contaminating falsity will leak out. Now the standard approach to negation, beloved of logicians and sticklers alike, says that two negations are similar to no negations. But isn't that like saying two wrongs make a right? That using both a belt and suspenders will combine to make your pants fall down? Or that using two condoms is just like having unprotected sex? Is that not unintuitive? And do you not doubt that the principle of prophylactic over-negation is not more reasonable?


  1. Pflaumbaum said,

    January 26, 2012 @ 3:43 am

    It would be interesting to know whether I don't doubt… is used particularly often in a concessive way: I don't doubt that X; but Y. Three of your five examples are, and a significant number of the examples on COCA are followed by but, I just, it's just etc.

    If so, might it be that it's not that is primarily meant to reinforce the concession, not the negativity of the I don't doubt clause as such?

    Absolutely. But if negation is being used to reinforce rather than deny, then that would seem to support my point. Consider an alternative concessive formula: "I suppose he might be right that…". Now if we negate that (e.g. "I don't suppose he might be right that it's going to snow") we get something that is (a) a little odd (though it get's better as "I don't suppose for one moment that he might be right that it's going to snow"), and (b) is not inherently concessive at all, unless interpreted as a negative question.
    – dib

  2. John Walden said,

    January 26, 2012 @ 3:59 am

    It's not unlike the spraying around of more auxiliaries than are normally used in order to swaddle the request in as much civility as possible:

    I would be grateful if you would tell me……

    If the ladies and gentlemen will accompany me, I will show them to their table.

    Is it fair to say that in an arm-wrestle between exquisite politesse and English's tendency to only do things once, redundancy is the winner?

  3. Lukys said,

    January 26, 2012 @ 5:47 am

    It actually took me a second reading with the "don't" removed to make me realise why there was an error. I really don't know whether that's my tiredness or the fact that it could be an easy mistake to make because it sounds natural upon hearing.

  4. PaulB said,

    January 26, 2012 @ 5:56 am

    Negative concord seems natural enough to me, but so does litotes. Can you have prophylactic over-negation as well and still understand what stuff means?

  5. Faldone said,

    January 26, 2012 @ 7:44 am

    I'm with Pflaumbaum, at least in examples 1 and 5. Another reading might be "I don't doubt that Kevin is low enough to … but that's not the point. I just don't think that he's clever enough" for 1 or "I don't doubt that the books are good (or …), but there are other problems that prevent me from finishing them." for 5. 4 seems like it needs more context for us to comment on it and 3 just seems wrong. 2? Well, all of these can be a little difficult to parse, but I think 2 is just plain wrong, too.

    If I understand you right, then I considered something similar. Could "It's not that I don't doubt the sincerity of their desire to protect the talent" be understood as "The issue here is not that I don't doubt the sincerity of their desire to protect the talent (although in fact I don't doubt their sincerity)". My conclusion was that while in principle this formulation could be used in such a way, the surrounding text didn't make that interpretation available in any of these cases. In the case of the Mann interview, it could just be that the article doesn't include relevant additional comments he made. Even if that were the case, however, I think my point would still stand, except applying to the reporter/editorial staff at AP and not to Mann himself.

    I hope nobody comes up with that tired "I couldn't fail to disagree with you less" chestnut, because I don't like to have to hurt nobody without they ain't no chance of no gold in it for me.

  6. Ran Ari-Gur said,

    January 26, 2012 @ 8:13 am

    @John Walden: I don't think that has anything to do with politeness. When "if" is used to propose an agreement, the same auxiliary is frequently used in both clauses:

    > I won't tell him if you won't.
    > I'll go if you will come back.


  7. Rube said,

    January 26, 2012 @ 9:33 am

    Is it possible that "I don't doubt" and "It's not that I don't believe" are getting mashed up between brain and tongue (or typing fingers)?

    In a lot of these constructions, it would work to say, for example, either "It's not that I don't believe some of the books are good" or "It's not that I doubt some of the books are good."
    I think it's pretty easy to get mixed up and try to say both at the same time.

  8. Glen Gordon said,

    January 26, 2012 @ 10:38 am

    When I think on it, I wonder if it isn't some sort of "non-emphatic emphasis" by way of stretching and padding the negation with a plethora of unnecessary verbage. This way, it gives the person being addressed the extra time to fully process the negation as it's happening so as to not mistake it as anything else.

    This phrase-stretching in itself is a kind of emphasis, isn't it? So instead of simply "I don't doubt", it becomes "it's not that I doubt" whereby the negation has been ripped from its normal position and the rest of the sentence has been subordinated to a relative clause in order to prop up the negation as theme. Reminds me of other stretchings like "It is he who has made it work." instead of simply "He made it work.

  9. jmmcd said,

    January 26, 2012 @ 12:53 pm

    Part of the issue is that "doubt" can be used in two ways:

    "I doubt X" — I suspect X is false

    "I doubt X" — I suspect X is true []

  10. Eric P Smith said,

    January 26, 2012 @ 6:07 pm

    @jmmcd: doubt may be ambiguous when it has a clause as direct object, as in “I doubt that he will lose”. But it’s unambiguous when used with a NP as direct object, as in doubting or not doubting one’s sincerity, or the power of colloidal silver.

    Multiple negations are very fashionable in English, and especially fashionable in polite turns of phrase. Accordingly, I wonder whether the phrase “It’s not that I don’t doubt his sincerity” is processed by some people (consciously or unconsciously) as follows?

    • The speaker used a polite word, sincerity, rather than an impolite word like dishonesty;
    • The speaker used lots of negatives and so he is probably trying to be polite;
    • It is therefore inconceivable that he may be accusing anyone of anything bad.

  11. John Walden said,

    January 27, 2012 @ 4:00 am

    @Ran Ari-Gur. I don't want to derail the topic but the fact that more auxiliaries are used in proposing or negotiating conditions than in threatening ones might indeed have something to do with politeness. Nothing usually happens for no reason.

    Sheer number of words does often have something to do with appearing conciliatory, a statement the possibility of whose truth I don't doubt that you daresay wouldn't mind being prepared to consider admitting.

  12. Eric P Smith said,

    January 27, 2012 @ 5:38 am

    @John Walden. I agree that more auxiliaries are used in “swaddling requests in civility” and in polite negotiation than in threats, but I don’t think it’s any old auxiliaries. I think it’s shall, will and may together with their preterites should, would and might. I’m not sure that the same holds of the auxiliaries is, have, must or ought, or should in the sense of 'ought to'.

    And I think it’s predominantly the preterites should, would and might. Personally, running against the fashion, I prefer to use the present tenses, precisely because I find them more direct: I think I always use “I shall be glad if you will…” rather than “I should/would be glad if you would…”, in both speech and writing.

  13. John Walden said,

    January 27, 2012 @ 9:03 am

    @Eric P Smith. Fair point. Linking it up with the thread, there is also the use of negatives to soften the blow of suggestions, without reversing the meaning:

    Should we be going?

    Shouldn't we be going?

    With "not" I think it's more swaddled. That also seems doable with:

    "Oughtn't we (to) be going?

RSS feed for comments on this post