Saturday, March 06, 2010

Negating must and have to

In most ESL grammars, you find claims that must and have to result in different meanings when negated even though their affirmative meanings are essentially the same. That is, negate must and you end up with a prohibition. Negate have to, on the other hand, and you get a lack of obligation.

It just occurred to me that this has less to do with must and have to and more to do with the scope of negation. I'm not claiming this as a new discovery. I've probably read or heard it before and simply forgotten, but it just struck me now and I don't think I'll forget again.
With must, negation comes after must, or is part of it in the case of mustn't. Periphrastic have to, though typically takes its negation to the left and requires do support. Here are two examples:
  • You mustn't do that.
  • You don't have to do that.
What I just noticed was that, you also have the option to move the not to the right of have to. In so doing, you change the scope of negation and end up with the same meaning as mustn't.
  • You have to not do that.
Attested examples include:
  • The President has said whatever is done has to not add to the deficit. 
  • If Mitt Romney had to not borrow lines, his, his rallies would have been silent.
  • I think you have to not go in there with your hair on fire yelling and screaming.
These are, admittedly relatively rare (at least two orders of magnitude less common than not have to), but I think they're fully grammatical.

13 comments:

  1. That's completely true about have to, as well as all other main verbs (to use the traditional term), but the original claim still seems like a meaningful statement about must, in that the various auxiliary verbs behave differently. Some behave almost like main verbs, preferring to give their "not" a high attachment but also allowing a low one, though without physically moving the "not" to indicate its scope ("I have not been there" vs. "I have been there some weekends and I have not been there other weekends"; "I cannot go right now" vs. "I can go right now or I can not go right now, whichever is more convenient for you"); others seem perfectly neutral in high vs. low attachment, with either being perfectly natural depending on context ("no, you may not touch it!" vs. "well, you may not agree with me, but . . ."); and a few, including must, seem to allow *only* low attachment, with no way to get high attachment short of creating a wrapper clause ("it's not the case that you must go" vs. "you must go"; "it's not the case that he might say it" vs. "he might not say it").

    ReplyDelete
  2. Andrew3:32 AM

    But then why does "You can't do that" mean "You aren't allowed to do that" or "You aren't able to do that" rather than "you are able to not do that"?

    ReplyDelete
  3. Yes, it seems I've not considered the potential variety. I'll blame it on much marking!

    ReplyDelete
  4. Circeus11:39 AM

    Andrew: I would offer up that auxiliary cannot have negation to the left as a possible explanation.

    ReplyDelete
  5. I'd like to add that must is part of a class of modals that can not take negation (or non-negatable modals). There is a left-to-right syntactic hierarchy in English. The negative marker 'not' generally subordinates everything to its immediate right. Must is of a category of modal that cannot accept such negation (the only thing it can accept is certain adverbial modifiers such as 'really').

    Have is of an entirely different structural class of modal which can accept negation (because the aspectual auxiliary is part of it's structure and aspectual auxiliaries can always be fully declined for person, number, tense, and also negated). This class is fully versatile in the same way that the non-modal usage of that verb would be.

    So with have you have more options:

    I have to go.
    (opposite modality)
    I don't have to go.

    I have to go.
    (same modality for opposite assertion)
    I have to not go.

    ---------------------------------

    Mustn't is actually a false contraction as it combines two separate syntactic and semantic elements -- the modal 'must' and the negative marker 'not' from the phrase the modal subordinates.

    So that "I mustn't go to work" is actually "I must [not go to work]" and not "I mustn't [go to work."

    There is no opposite mood to must.

    ----------------------------------
    BTW, just a point of contention, there is no such modal as "have to" (or going to, need to, ought to). Rather "have" as a modal requires that its subordinate take the infinitive form (must requires the finite form).

    ReplyDelete
  6. Actually, it occurs to me that there are some restricted circumstances in which mustn't has a high-attaching -n't; searching Google Books for "no, there mustn't", I find examples like these:

    > The key is to give up on the idea that, […], there must be […]. No, there mustn't. (More accurately: no, there needn't.)

    > "But there has to be a way!" I said. "There must be. She's a genius." ¶ "No, there mustn't," said Brendan Regan flatly. "Life is not like a story. […]"

    Need, incidentally, is interesting in that it's the opposite of how Drew Ward describes must: it, like dare and durst, is a modal that basically requires negation. High-attaching negation. And as my first link suggests, needn't seems to function as the "opposite mood to must" that, as Drew Ward says, does not normally exist otherwise.

    ReplyDelete
  7. Ran,

    I wouldn't say that the presence of examples of usage necessarily means that usage is grammatically or semantically valid. I hear mustn't from British colleagues all the time, but the message they are expressing is must not... Instead of (anti)must.

    The examples you've pulled from google books do appear to be expressing a counter mood, but I'd be curious what the dates were as I'm not sure that usage would be found in contemporary texts.

    As for dare and need, they could certainly be used in the affirmative. "Do you need to go?". "Does he dare do it?"

    ReplyDelete
  8. > I wouldn't say that the presence of examples of usage necessarily means that usage is grammatically or semantically valid.

    Oh, definitely. But what happened here is, I thought of a very narrow context where I think "mustn't" can validly mean "needn't" (viz. the context immediately after a "must" assertion, that is now being contradicted), and then I crafted a Google-search to find real-life examples of that context, rather than trying to make up a hypothetical example to demonstrate it.

    > I hear mustn't from British colleagues all the time, but the message they are expressing is must not... Instead of (anti)must.

    Yes, 100% agreed. (anti)must is incredibly rare, and I didn't mean to suggest otherwise. (Even with the crafted search I used to find exactly this usage, only three of the first ten hits were using it this way.)

    > The examples you've pulled from google books do appear to be expressing a counter mood, but I'd be curious what the dates were as I'm not sure that usage would be found in contemporary texts.

    If you click the links in my above comment, you'll have as much information about them as I do. They both purport to be from within the past ten years, and they do seem recent to me, but Google Books' metadata can sometimes be sneakily deceptive.

    > As for dare and need, they could certainly be used in the affirmative. "Do you need to go?". "Does he dare do it?"

    I think you mean "Need that happen?" and "Dare he do it?"; but yes, you're right: they only require negative polarity, not necessarily actual negation.

    ReplyDelete
  9. Andrew2:28 AM

    I think part of the explanation underlying "there mustn't" is that "there must be" and "there mustn't be" don't express obligation at all, but rather probability. Compare "I must be on time" to "there must be a solution." There's no subject in the existential expression "there is" to whom an obligation attaches.

    But that's not the whole story, since it seems to me that [A: "There must be a solution." B: "No, there mustn't."] and [A: "There must be a solution." B: "No, there must not be a solution."] have different meanings. I think here's a case where the context, tone and emphasis make a big difference. Here, perhaps, the difference lies in whether B places the emphasis on simply contradicting what A has just said (in the first example) versus substituting an alternative opinion (in the second opinion).

    ReplyDelete
  10. @Andrew: It's true that I chose a search that produced epistemic examples (where must denotes probability) rather than deontic ones (where it denotes obligation); but with some effort, I think similar-but-deontic examples can be found as well. Try searching for "you must * * * no I mustn't"; there's no one example that I can point to as absolutely unambiguous, but there are plenty of cites that seem to be doing the same thing.

    > I think here's a case where the context, tone and emphasis make a big difference. Here, perhaps, the difference lies in whether B places the emphasis on simply contradicting what A has just said (in the first example) versus substituting an alternative opinion (in the second opinion).

    I agree completely.

    ReplyDelete
  11. Hey Ran,

    Could you send me an email, I'd like to learn more about how you customize your google searches as I've not had the best luck trying that myself.

    drew (dot) ward (at) calleteach (dot) org

    @Andrew I think context is as you say a primary point. English a bit unique in that it uses it's various parts of speech to express a great variety of meanings. Consider that we only have (I forget the exact count) I think 27 prepositions and each of those can have up to 7 different prepositional uses. Modals are the same way. They can be separated into distinct structural classes which stay the same. However each modal can often express multiple moods. And, there are a few that seem to override their constraints when used to express certain moods (really odd as it's rare that a semantic change overrides syntax).

    This is a really neat discussion as modality is one of the most researched yet still least understood areas of language.

    ReplyDelete
  12. It is indeed a matter of scope of negation; it is also known as internal/external negation. See CGEL pp.175-76; look under 'negation and modality' in the Conceptual Index for more.

    ReplyDelete
  13. Anonymous9:13 PM

    Nice Post. This record helped me in my college assignment. Thanks Alot

    ReplyDelete