(Not) Underestimating the Irish Famine

« previous post | next post »

Breffni O'Rourke writes:

Here's one for the 'cannot underestimate' files. The publicity material for the recently published Atlas of the Great Irish Famine (which may coincide with the printed blurb or the preface; I haven't been able to check) starts off with (variants of) this:

The Great Irish Famine is the most pivotal event in modern Irish history, with implications that cannot be underestimated.

From the book's website:

The Great Irish Famine is possibly the most pivotal event/experience in modern Irish history. Its global reach and implications cannot be underestimated.

From a feature article in the Irish Times by one of the book's editors:

The Great Irish Famine is possibly the most pivotal event in modern Irish history and its global reach and implications cannot be underestimated.

The 'cannot' = 'must not' interpretation ('Don't underestimate the implications of the Famine') certainly doesn't work in this context. It would be strangely defensive to introduce a landmark work of scholarship by mentioning doubts about its significance, even if only in dismissing them.

The "cannot understimate" files are a subset of our misnegation files. If you're new to the discussion, you might want to start here.

Update — An interesting trend is visible in the Google Books ngram corpus:

Based on a spot check, nearly all of the "X cannot be underestimated" examples are either ways to say that X is important and should be recognized as such, or are part of a discussion of whether this way of phrasing that thought is correct or not.

Some related (?) trends:


  1. Eric R-S said,

    September 16, 2012 @ 12:46 pm

    Perhaps it's a mistaken substitution for "cannot be understated"?

    [(myl) But "cannot be understated" has exactly the same problem: it would mean that the implications are so small that no estimate could be too low.]

  2. G said,

    September 16, 2012 @ 1:00 pm

    Isn't it possible that the writer has adopted the "cannot" = "must not" interpretation, but uses it as a stock phrase to express the topic's importance without considering that it brings up doubts, if only rhetorically?

    [(myl) As discussed at tedious length here, the equivalence of "cannot" and "must not" involves a reversal of the implied scope of the negative and the modal,

    ¬PA → O¬A

    which works only in the case of deontic modals (those dealing with permission and obligation): if it's not permitted that X, then it's obligatory that not X. This doesn't apply in cases where can means "to have the power, ability or capacity" or "to be a possible contingency". Thus "You cannot shoot people who annoy you" may be interpreted to mean that "You must not shoot people who annoy you"; but the proverb "An empty sack cannot stand upright" does not mean that "An empty sack must not stand upright".

    Expressions like "we cannot under/overestimate" seem to me to be about capacity and not permission. The logically reasonable expression, "We cannot overestimate X" means that X is so big that no matter how we try, we're not able to assign it too high a value. The relevant deontic interpretation of "we cannot underestimate X" would be "we are not permitted to underestimate X".

    Breffni's point was that even if the modal is a deontic one here, the usage would still be rhetorically odd. I'm not entirely convinced — but I agree that the authors probably didn't mean that it's ethically impermissible to underestimate the implications of the Irish Famine, ]

  3. Bloix said,

    September 16, 2012 @ 1:29 pm

    There are uses of "cannot" that imply "without adverse consequences."
    "You cannot drive in that HOV lane because you don't have three people in the car" means something entirely different from "you cannot drive on that road because it is flooded out." The first means, "What you propose is illegal and may result in a ticket." The second means, "what you propose is an impossibility,"

    You are reading "cannot be underestimated" in the second way – "it is impossible as to underestimate the reach and implications of the Famine." Let me propose that you read it the first way: the Irish Famine's "global reach and implications cannot be underestimated [if you expect to understand modern Irish history]." That would be a meaningful sentence.

    [(myl) You may be right. But I would be more convinced of this view if we didn't so often see similar expressions that are not amenable to a similar analysis: "It is difficult to underestimate the importance of X", "It is impossible to underestimate the implications of Y", etc. Things like those don't work in your HOV Lane example: "It is difficult to drive in that HOV lane because you don't have three people in the car", etc.]

    If the sentence had been written to be literally logical and complete, it would be clearly be untrue: the Irish Famine's "global reach and implications cannot be overestimated." Well, of course they can be. "The global reach and implications of the Irish Famine far outstrip those of the Second World War."

    And you can't really supply an implied meaning that fixes the problem. "Cannot be overestimated [by anyone who expects to understand Irish history]"? That's almost the opposite of what the author means to say.

    So a "correct" version of the sentence would have to be understood as an exaggeration that is not meant to be taken literally, while the "incorrect" version has a grammatically correct and accurate meaning once the implied condition is supplied.

  4. Jeroen Mostert said,

    September 16, 2012 @ 1:33 pm

    "The 'cannot' = 'must not' interpretation ('Don't underestimate the implications of the Famine') certainly doesn't work in this context."

    Sure it does. The author is telling you, in no uncertain terms, that it would be improper to underestimate the implications of the Famine, because the Famine is bloody important and you'd best remember that for your next History test.

    "It would be strangely defensive to introduce a landmark work of scholarship by mentioning doubts about its significance, even if only in dismissing them."

    The author is not mentioning doubts about the significance of the *work*, or even about the significance of its *subject matter*, but is explaining to you that the Famine was big. Really big. You just won't believe how vastly hugely mindbogglingly big it is… so that's why we're telling you that its implications should not be underestimated, before we delve into more detail as to why not.

    Stylistically, this is a rather poor way of starting off if you're not Douglas Adams, but it's not *strange*, and taking it as counter-evidence of the "deontology light" theory (which I'm a total fan of, by the way) is certainly going too far. The fact that people saw fit to copy this no less than twice in slightly different words shows me that there isn't exactly a big heap of sparkling literary wit that's being tapped here — in fairness, you wouldn't expect that of blurbs.

  5. Jeroen Mostert said,

    September 16, 2012 @ 1:48 pm

    Having read the copious background discussion, what I wonder is this: how would one go about validating the competing theories for explaining this use of language? That is, how would you decide if the "poor monkey brains" theory is correct, which in a nutshell says that this is just wrong but we fail to see it as such because we can't not properly handle all those negatives, versus the "deontology light" theory, which says that these expressions are semantically correct because the statement that one cannot do X carries an implicit "because it would be bad to do X" connotation?

    My money's on the "deontology light" theory, but only because I think it's so dang plausible. And also because its consequences are so elegant: "this can't be underestimated" and "this can't be overestimated" can effectively mean the same thing, but only because they *don't quite mean the same thing*. "This can't be underestimated" means "this should not be underestimated [because I, the author, along with all right-thinking people of this world, would think considerably less of you if you did]" but "this can't be overestimated" means "this can't be overestimated [because any attempt to do so is foiled by the gob-smacking enormity of the thing]".

    [(myl) Again, this would be more persuasive if it weren't for the high frequency of apparently similar cases to which the "deontic modal" story doesn't apply: "It is impossible/difficult/hard to underestimate/understate/undervalue X", and so on; and the relatively low frequency (in the same meaning) of cases of the form "may not underestimate/understate/undervalue" (which is the prescriptively correct way to express deontic modality), and similarly the low frequency of other explictly deontic paraphrases like "it is forbidden to underestimate" or "we are not allowed to underestimate" or whatever (if indeed these occur at all). ]

    However, deciding a theory is right because it's appealing is not exactly scientific, and (almost needless to say, I think) I'm not a linguist, lacking both proper papers and street cred. So how would you settle the matter linguistically, with all proper decorum? Questionnaires? EEGs? Sensory deprivation followed by a rapid-fire interrogation under hot lights?

    Incidentally, I'm not proposing *all* instances of misnegation can be correctified[*] — you can't fail to miss the difference between cases like "you can't fail to miss the difference" and "I am not saying that there's no circumstances where we would not use force". The semantic muddle-headedness of the latter is a prime candidate for the "poor monkey brain" theory.

    [*]"Correctified" isn't a word; like I said, papers nor street cred.

  6. Eric R-S said,

    September 16, 2012 @ 1:59 pm


    Rgh, I made a similar mistake myself! :-) I meant to suggest that the writer made an inadvertent substitution for "cannot be overstated".

  7. Jeroen Mostert said,

    September 16, 2012 @ 2:44 pm

    @myl: not being a linguist, I don't know how the deontic modal works, exactly (or even how it would taste on a sandwich), but I do think there's room for an extension somewhere.

    You're quite right when you point out that "it's forbidden to underestimate" would be extremely odd, and in fact I think this is very relevant. "You can't underestimate" *cannot mean* "you are [morally/ethically/in-the-author's-opinionally] forbidden to underestimate", it can only mean "I think you ought not underestimate [because bad consequences are attached to it]". This is, I don't know, the no-right-thinking-mind modal, the ought-not modal.

    The ought-not modal *does* apply to "it's impossible to underestimate". This does not mean "it is logically/physically impossible to underestimate" (unless you're being contrary), but "it's *unthinkable* that anyone would underestimate". It's a valuation, not a logical predicate. "It's impossible to condone murder" obviously doesn't mean there is some sort of fail-safe in our brains that will prevent us from attaching positive assent to acts of homicide towards our fellow human beings, it means we really ought not condone it if we know what's good for us. "It's hard to underestimate it" means "if you are in a right frame of mind, it ought to be difficult for you to underestimate it" and not "because of its extremely limited impact, you could almost always think a little less of it". The latter is a plausible reading of the meaning going by the rules, but it's so far from the usual meaning that it veers into comedy if you insist on it: "it's hard to underestimate this man's talents, because he really has none to speak of".

    Obviously this is (as I've presented it) a just-so story on no proper linguistic footing whatsoever, hence my question that I'd be really interested to know how we'd go about settling between the theories.

    [(myl) But this move doesn't seem to me to help, since "If you are in a right frame of mind, it ought to be hard to underestimate X" has exactly the same problem that "it's hard to underestimate X" has.

    Perhaps it will help to clarify the difficulty if we expand "underestimate" to "assign too low a value to". Then the phrase becomes "It's hard to assign too low a value to the implications of the Irish Famine". Can you maintain the same interpretation in that case? ]

  8. Dick Margulis said,

    September 16, 2012 @ 2:46 pm

    All that aside, what the heck does "most pivotal event" mean? Either an event is pivotal or it is a burp on the way to or from the pivotal event.

    [(myl) But the OED glosses the relevant sense of pivotal as "central, crucial, vital", and things can certainly be more or less central, more or less crucial, more or less vital…]

  9. ALEX MCCRAE said,

    September 16, 2012 @ 3:46 pm

    Not directly related to this, in effect, use of a double negation to unwittingly create a 'net positive' statement, but kind of in a similar vein, is the use of "arguably", and "inarguably".

    I admit, at one time I had a hard time differentiating between the proper use of the two related words, but have come to understand that "inarguably", for one, means that there is no room for argument, in say the assertion that "Serena Williams is, inarguably, the most dominant women's player in tennis, today."

    Whereas. if one asserts that the current number-one ranked female tennis player in the world today, "is, arguably, Victoria Azarenka", the use of "arguably", in this instance, implies that Ms. Azarenka may be presently officially ranked as the world's top women's player, but some might 'argue' a strong case for Serena Williams as the bona fide 'numero uno', on the sheer merits of her recent string of dominating major victories—Wimbeldon, Olympic gold, The U.S. Open.

    In this case, "arguably" suggests room for debate, and that Azarenka's top-spot is open to questioning. (She's still a heck of a player, and a breath of fresh air on the women's tour.)

  10. Ø said,

    September 16, 2012 @ 4:12 pm

    In statements of this kind involving "arguably" and "inarguably", I do not take the two words to have directly opposite meanings. To me "X is inarguably the top player" means that the statement "X is the top player" cannot be argued *against*. Whereas "X is arguably the top player" means that one can argue *for* the statement "X is the top player".

  11. Jeroen Mostert said,

    September 16, 2012 @ 4:46 pm

    @myl: Yes. "It's hard to assign too low a value to the implications of the Irish Famine" crucially does not say *why* it is hard to assign too low a value. The reason it's hard is not because the implications are of so low a value that we would have trouble getting the value low enough (indeed, this interpretation is forced exactly *because* it is almost mathematically unsound to say this, and why "it's hard to overestimate" is hyperbole), but because it would be hard for any reasonable person to do so. Of course the question now becomes whether or not we are justified in assigning this meaning and not another, because the author doesn't *state* which is the case. The argument now hinges on what we are saying — that it is *hard* to assign too low a value (i.e. it is unreasonable to do so), or that it is hard to assign *too low* a value (i.e. we could always go lower)?

    Although it's true that "underestimate" makes the former interpretation more obvious because the "too low" part is hidden in the verb, it still does not at all seem unreasonable to me to extend the "ought not" interpretation in this case too. It is true that if an author chose to phrase it this way, you would expect immediate follow-up to clarify precisely *why* the author thinks this, but that's only because "it's hard to assign too low a value" from its texture and feel implies a sort of mathematical precision that "it's hard to underestimate" does not.

    It's also true that, if you stretch far enough, eventually the interpretation flips in favor of its opposite number. "It would be hard for me to hold this man in lower esteem than I do" *could* mean "because I hold this man in such obviously high esteem that I couldn't possibly lower my standards", but would more likely be taken to mean "I think he's pond scum and there's not much he could do to make it worse". The ambiguity is exploitable by a sufficiently devious speaker, though.

    In cases such as "it's hard to underestimate", however, I wouldn't go for "this is wrong" or even "this is ambiguous" but flat-out "it means that we ought not underestimate".

  12. Andy Averill said,

    September 16, 2012 @ 5:33 pm

    But in any event, you rarely see "cannot be underestimated" used in the sense "is really really small". One generally doesn't find oneself writing about how unimportant something is, except I guess just to be funny.

    And in fact, the first n results from a Google search for "cannot be underestimated" (leaving out meta-examples such as other Language Log columns and articles by Ben Zimmer and William Safire) all seem to run in the same direction as today's example:

    The complexity of bank reform cannot be underestimated.

    Brailsford's influence over British cycling cannot be underestimated.

    The damage wreaked upon politics by the MPs' expenses scandal cannot be underestimated.

    Iranian cyberthreat cannot be underestimated [headline]

    The importance of companion care cannot be underestimated

    Influence of Queen's reign cannot be underestimated

  13. Jerry Friedman said,

    September 16, 2012 @ 6:05 pm

    Here's a clearly deontic one, if it makes a difference. "Clearly" to me, anyway.

    Manchester United ace Wayne Rooney says England cannot underestimate Italy.

    England won their Euros group after victory over Ukraine last night.

    Rooney said: "People say we've avoided Spain but I don't want to underestimate the Italians.

  14. Viseguy said,

    September 16, 2012 @ 6:14 pm

    I was thinking, wouldn't it be nice if the computer could flag a likely misnegation and ask, "Do you really mean to say this?". Then I read "No post too obscure to escape notice" (http://languagelog.ldc.upenn.edu/nll/?p=1925) and realized that it's impossible to understate the probability of this actually happening.

    Computer: Do you really mean to say this?
    Me: Absolutely! I thought it through — there's no misnegation there!
    Computer: Could you rephrase that without using a double negative? But seriously … isn't it "[]possible" to state that "the probability of this actually happening" is "0"?
    Me: Well, sure.
    Computer: And wouldn't that be understating "the probability of this actually happening"?
    Me: Um, er…
    Computer: Especially given that I said what I just said.
    Me: Okay, okay. It would appear that "the probability of this actually happening" is "1".
    Computer: Yes, it would appear so. Would you like to rephrase your original statement to "… it's impossible to overstate the probability of this actually happening"?
    Me: Wouldn't it be "possible" to say that "the probability" is "1.5"?
    Computer: Fucking idiot!

  15. GeorgeW said,

    September 16, 2012 @ 6:54 pm

    Which of the competing theories would better explain why no (or almost no) native speaker of English would misunderstand the author's intent? Was anyone confused?

  16. L said,

    September 16, 2012 @ 7:07 pm


    You signon ID would explain that!

    But seriously, English speakers are pretty good at communicating with speakers of broken English, regional English, oddball English. I think English might be especially flexible that way. It's very fault-tolerant.

  17. GeorgeW said,

    September 16, 2012 @ 7:17 pm

    @L: I don't think I would put this in the same category as broken or oddball English.

  18. Pharmamom said,

    September 16, 2012 @ 8:07 pm

    @ L. I find your saying that English is very tolerant of variation interesting. I spent three weeks this summer in Italy, speaking Italian; and I also have a very good friend here who is Italian. In my experience, I can understand my friend's attempts at English, and I understood many Italians' efforts, however poor, far better than they understood/understand my almost fluent Italian.

    If I use the incorrect tense of a verb in Italian, confusion ensues. In contrast, my friend knows pretty much only the present and some bits of proximate past in English, but I am damn good at figuring out what she means to say. I thought this difference was due to my verbal flexibility, but perhaps it is more the nature of the languages. English is remarkably fluid, isn't it? That's what drives prescriptivists batty. But other languages have rules that, evidently, no one native ever breaks. Who knew?

    A great example of this occurred when I told my friend, in English, that we were having fish for dinner. She was flummoxed by using "to have" as a proxy for "to eat.". If I had said "noi abbiamo del pesce a cena," she would have had no idea what I was saying. It doesn't even make sense to ME in Italian. Weird.

    My working theory is English's flexibility is what makes word play in our language so easy and so damn funny–and common. Has LL ever looked at verbal play profiles of languages?

  19. bloix said,

    September 16, 2012 @ 9:44 pm

    The all-knowing Google reveals that Ben Zimmer, writing in the NYT in 2011 (in an article quoting you) put forward a theory of "cannot be underestimated" that is close to my proposal. He says, "Instead of thinking of cannot as indicating the impossibility of an action, we can instead treat it as pointing to our moral obligation not to do something — namely, underestimating something that we shouldn’t underestimate."


    [(myl) Yes, that's the explanation (2) of the four possibilities suggested here:

    1. Our poor monkey brains just can't deal with complex combinations of certain logical operators;
    2. The connection between English and modal logic may involve some unexpected ambiguities;
    3. Negative concord is alive and well in English (or in UG);
    4. Odd things become idioms or at least verbal habits ("could care less"; "fail to miss"; "still unpacked").

    I remain skeptical of explanations of type (2) for these cases, partly because of what the phrases with "cannot" seem to me to mean, but mostly because of the similar examples with "impossible", "difficult", "hard" etc. Jeroen Mostert (in the comments above) is confident that they are somehow the same sort of modality, but I don't see how to make that work, while freely admitting that someone who knows more than I do about formal pragmatics might be able to put his intuition on a sound footing.]

  20. etv13 said,

    September 16, 2012 @ 11:09 pm

    These misnegation posts make my head hurt.

  21. Jerry Friedman said,

    September 17, 2012 @ 12:01 am

    @L and Pharmamom: MYL posted on the subject of misnegation in other languages here.

    I wonder whether any work has been done on differences in "fault tolerance" between languages. I don't even know how you'd measure it.

    @MYL: Are examples where "underestimated" is conjoined with another verb relevant?

    "In choosing the title North-South: a programme for survival, the Commission indicate that the question of justice for the world's poor cannot be underestimated, nor can it be put off. There needs to be further discussion over the implementation of some of the report's recommendations, but without the political will to solve the problem there is little hope of progress." Can North and South be friends?

    "…that importance cannot be underestimated or undervalued without there being consequences, however subtle, to the collective force and effectiveness of your business." Every Relationship Matters.

    "A painful lesson of the Nazi era is that the irrational cannot be underestimated or neglected." Sources of European History Since 1900.

    These look to me like deontic uses where the authors chose the word they wanted, not errors for "overestimate". I'm not arguing that every instance of "cannot [be] underestimate[d]" is of this type, since I don't see how that would explain your examples with "it is difficult to underestimate". Maybe tellingly, Google Books has only one hit for "it is difficult to underestimate or", which I can only read as a misnegation.

    "…it is difficult to underestimate – or to overstate – the importance it had in its own day for encouraging other black writers to speak for the black child." Virginia Hamilton.

    [(myl) Some of these suggest pretty strongly that their authors are using "cannot be" to mean "should not be" in these constructions.]

  22. Jerry Friedman said,

    September 17, 2012 @ 12:03 am

    Actually, I think the business one isn't deontic, but it's too late at night.

  23. Nibbles: Irish Famine book, Breeding for adaptation, Neolithic diets, Randy Thaman, Ecological Babylon, IPR for smallholders, Botanical gardens said,

    September 17, 2012 @ 3:40 am

    […] underestimate the importance of a new book on the Irish Famine, despite the weird construction used in praising […]

  24. Mark F. said,

    September 17, 2012 @ 7:28 am

    I wonder why "cannot be overestimated" sounds wrong. I find myself wanting to use "cannot be overstated" instead.

    [(myl) From the Google Books ngram corpus:

    And this:

    Finally this:

    Freaky. The difference in behavior between "cannot be under/overestimated" and "cannot be under/overstated" suggests that there's some idiom-formation going on.]

  25. GeorgeW said,

    September 17, 2012 @ 7:37 am

    The words "pivotal" and "implications" might prime the reader to a certain interpretation of "cannot be underestimated" without doing hard work of logical parsing. (This would, I guess, be the monkey-brain theory).

    But, what about the writer who chooses the words? Or, the editor who presumably carefully looks for confusing statements? I don't think the monkey-brain theory works as well.

  26. L said,

    September 17, 2012 @ 8:35 am

    @ Jerry Freidman:

    So does this meant that misnegation is, or is not, somehow different (as George W said) from other fractures?

    Let us assume that it is different. It might yet be a monkey-brain matter, but in a whole other sense. I'm asking here, I'm not asserting anything: could it be that, um, the following? When monkeys (or probably closer, great apes) use a specific call – say, danger – do they ever use it as its own negation – say, all-clear? If so, it might point to a "monkey-brain" phenomenon far deeper than human language; but not a monkey-brain limitation as such, rather a monkey-brain capability to correctly disambiguate contradictory meanings of the same call vocalization. Now that would be cool.

    We too often look down on our simian cousins, while at the same they are looking down at us, preparing to throw poop. Who's smarter now?

    It also occurs that the "four theories" are not mutually exclusive. The notion that "only one thing causes X" is not always applicable – sometimes but not always.

  27. L said,

    September 17, 2012 @ 8:55 am

    @Jerry again, on another topic

    >I wonder whether any work has been done on differences in "fault
    > tolerance" between languages. I don't even know how you'd measure it.

    As I reflect on this, I suspect it isn't a function of language but more apt to be a function of exposure to variation. Of course, some languages may exhibit more variation than others, and I don't know how you'd measure that either (count synonyms? a proxy at best).

    About the same time I mis-learned that pun on Kafka, I was visiting Moscow and was advised to speak my broken Russian in touristy areas, but speak my (tolerably) good English when outside of them. I was advised that locals who encounter a lot of broken Russian (in foreigner-frequented areas) would find it easy to interact with my broken Russian and appreciate the respect of trying, but those who do not (in other areas) would struggle with it; whereas I would deal well with either the skilled or broken English that most of them would have learned in school. It's purely anecdotal, and it may no longer be relevant 20-something years after the Soviet Union fell, but it's very suggestive.

    But, and here's the kicker, it proved to be mostly wrong; locals out "in the neighborhoods" who felt like trying to work with my really-bad Russian had no problem whatever – we got our messages through easily enough, they got it somehow – and sometimes replied in no better English, and we made it work. Or sometimes we codeswitched freely, whatever seemed to help. It seemed to me to be a confidence issue more than anything else; those who tried, succeeded. Of course one can readily understand how, in the touristy places, that confidence might be increased more rapidly. After all, there is motivation to try and in fact, it turns out to work fairly well even for newbies.

    That says little or nothing about English or Russian as such, but something about cultures coming out of totalitarian isolation… or more likely out of isolation regardless of cause.

  28. George said,

    September 17, 2012 @ 9:00 am

    As a number of commenters have pointed out, "cannot be underestimated" is far more often used to mean "is really important" than "is really unimportant" and, crucially, is usually understood to mean what the writer intended it to mean. So we manage to overcome any logical problems and get the message.

    I'm reminded of the term "lowest common denominator", which logically suggests the opposite of what is almost always intended when it is used in a non-arithmetical context. But we still understand what is meant. Maybe it's the suggestive power of the word 'lowest' – in the 'low down and dirty' sense – or of the word 'common' – in the 'common as muck' sense – but the logical problem gets overlooked. Or at least it gets overlooked by native speakers. I've been at meetings with large numbers of non-native English speakeres where a native speaker would warn solemnly that we should not settle for some lowest common denominator consensus and I do sometimes wonder how well that particular message is understood.

  29. quixote said,

    September 17, 2012 @ 9:37 am

    Just to add a data point from a non-linguist speaker of English: My immediate understanding of the "cannot be underestimated" phrase was "Haha. The twerp just said the opposite of what he means. It's cannot be overstated, doofus!"

    The alternate reading of "must not be underestimated" didn't even occur to me until I read the comments.

    I think that plus the Goog ngrams are evidence of the erosion of language, comprehension, and the general end of civilization as we know it. I'm with etv13: it makes my head hurt.

  30. Andrew (not the same one) said,

    September 17, 2012 @ 10:01 am

    In any case, 'should not/must not be underestimated' would have a different meaning. You can say 'X should not be underestimated' when X is only moderately significant, but there's a temptation to treat it as utterly insignificant. This doesn't seem to be what is meant here; it's clearly meant to imply that the Irish Famine is very significant indeed.

  31. George said,

    September 17, 2012 @ 10:18 am

    @Andrew (not the same one). I don't think you're right that appropriate use of 'X should not be underestimated' is limited to "when X is only moderately significant, but there's a temptation to treat it as utterly insignificant". It's appropriate to use it when there's a temptation to treat X as less significant than it really it (which may mean treating is as quite significant but not as truly, madly, deeply significant).

  32. jan said,

    September 17, 2012 @ 10:48 am

    A political columnist has a book called, Molly Ivins Can't Say That, Can She?

    Nickolodeon had a children's show called "You can't do that on Television".

    Sometimes "can't" is used as a polite substitute for "shouldn't". But it can be confusing for kids. "You can't beat him up!" "I sure can!"

    [(myl) The examples that you cite just involve the common interpretation of can as "be permitted to". There's a prescriptivist belief that this meaning should be allowed only for "may", with "can" limited to meanings associated with ability; but can = "be permitted to" has been in regular use since the middle of the 19th century. The relevant notions of permission/obligation include legal constraints, organizational policies, social conventions, individual grants of permission, and so on.

    And by a general principle of modal logic, "not be permitted to" is equivalent to "be obliged not to". This is enough to yield a coherent interpretation of "cannot be underestimated", in the way that people normally seem to understand it, as long as we interpret "cannot" as meaning something like "is under a moral obligation not to". (Other notions of obligation generally don't work.)

    For a full explanation of the phenomena here, we need to extend this to cover things like "it's hard to underestimate". Jeroen Mostert has a story to tell about this, but I don't understand it, at least not yet.]

  33. L said,

    September 17, 2012 @ 12:32 pm

    > I'm reminded of the term "lowest common denominator", which
    > logically suggests the opposite of what is almost always intended
    > when it is used in a non-arithmetical context.

    My pet peeve is the common, usually political, use of "fuzzy math" to indicate "unreliable, unclearly-thought-through, and possibly incompetent math."

    Fuzzy math is as precise as any other; I won't go into it here, but it is – as is the closely-related "fuzzy logic." Done correctly, it as correct as any other; done incorrectly it as incorrect as any other. "Fuzzy" is not a perjorative term when applied to mathematics.

  34. Brett said,

    September 17, 2012 @ 12:38 pm

    @George: I have seen two native speakers of American English get into a very animated argument about what exactly "lowest common denominator" meant. For one of them, it was just a fixed idiom, metaphorically meaning something like, "the lowest or worst instance that will ever be encountered." The second person accepted that "denominator" had to be interpreted metaphorically, but he parsed "common" separately. To him, it meant, "the lowest or basest instance that is nevertheless not actually repugnant"; he took "common" to indicate that the quality of the phenomenon was such that (virtually) everyone would at least accept it, if not like it. Acceptance was something everyone had "in common."

    After hearing this argument, I stopped using the idiom entirely, since I realized that I didn't have a really clear idea of what other people necessarily meant by it, nor did I have a precise meaning of it in my own head. I think this is relevant to the "cannot be underestimated" situation as well. I imagine that there are some people for which this usage is deontic, but for most people it is simply an idiom-to-be, whose formation has been made possible by the monkey brains' tendency not to notice the overnegation.

  35. Andrew (not the same one) said,

    September 17, 2012 @ 12:49 pm

    George: Sorry, I was insufficiently clear. What I meant was that 'X should not be underestimated' can be used when X is only moderately significant, etc., and so it clearly does not mean 'X is tremendously significant' (though it's compatible with that). But that was what the writer in this case was trying to convey, so what they said was not equivalent to 'should not be underestimated'.

  36. Mark F. said,

    September 17, 2012 @ 1:18 pm

    People keep saying that we all understand what people mean when they say "X cannot be underestimated", but that's only true on a coarse level. We know they mean "X is large," but, judging by this discussion, we can't always tell whether they mean "be careful, X is pretty large", or "wow, X is so large I don't think I'd be able to exaggerate."

    I think what has happened is a combination of the two leading theories. MYL has pointed out unambiguous examples of misnegation that use words other than "cannot", so it seems likely that sometimes people have used "cannot" in just that monkey-brain way. But others have pointed out unambiguous cases where "X cannot be underestimated" clearly means "those who underestimate X do so at their peril". The presence of those use examples makes "cannot underestimate" much more accessible even in cases where the desired degree of emphasis favors the "cannot overstate" sort of interpretation.

  37. Vince said,

    September 17, 2012 @ 10:39 pm

    What if we interpret it as an instance of capacity and not deontic after all? "It is hard to underestimate", "It cannot be underestimated", etc. are all perfectly reasonable statements (within acceptable bounds of hyperbole) if the assumed premise is that the evidence is so flagrantly and heavily against underestimating it that it's practically impossible.

    Presumably this isn't the psycholinguistic logic behind it, but it's a valid interpretation as far as I can see.

  38. Michael Ranieri said,

    September 18, 2012 @ 12:37 am

    From my naive standpoint, there are two ¨thoughts¨ I use to understand or when I utter (most of) the misnegations I use myself, or commonly read or hear.
    ¨can¨ as a modal has several interpretations. If one lays them out on a spectrum (authority, ability, opportunity), the range is significant. I think others have pointed this out.
    ¨not¨ may also be interpretive. For example, if I say I am ¨not hungry¨, most people would conclude I am satiated. If I say my horse is ¨not white¨, most people cannot conclude what color my horse is, and certainly do not conclude it is black.

    Combining ¨can¨ and ¨not¨ in a thought with another negation can and does provoke more than one understanding. I think that´s just fine – and what English thinking cannot be undone by.

  39. Jeroen Mostert said,

    September 18, 2012 @ 2:24 am

    @myl: I think the modal logic thing is a red herring. It's simple, neat and wrong (to steal from Mencken), offering an explanation only in a limited absolute case. You write in your 2004 post:

    [And it doesn't seem to be the whole story even in the "cannot" or "impossible" cases, because when people say "we cannot understate the importance of X" (the phrase Wallraff is discussing), they seem to mean "we cannot overstate the importance of X" (because it is so great), not just "we're obliged not to understate the importance of X" (because it is not negligible).]

    Well, it's exactly the opposite. "We cannot understate the importance of X" does *not* mean "we cannot overstate the importance of X" — this is a legitimate phrase with the same intent but not the same meaning. The meaning of the original *is* just "we're obliged not to understate the importance of X" (and not merely because it is not negligible, but because it is in fact very big). But even this is misleading because "obliged" implies the kind of deontic categorization that is far too strong and causes you to dismiss explanations along these lines when they don't fit this model. I think Vince (and others) are on to something by saying that we should interpret these as statements of capacity, but I still think this is not the whole story.

    Perhaps it helps to think of statements like "it's hard to underestimate the importance of X" as "it is difficult to conceive of a desirable outcome if the importance of X were to be underestimated". "It's impossible to underestimate the importance of X" then becomes "it is impossible to conceive of a desirable outcome if the importance of X were to be underestimated". The speaker is not saying the act of underestimating is itself inherently hard (or impossible) due to the nature of X, but merely that it ought to be so because no good can come of it. Does anyone want to take a stab at probabilistic deontic logic? :-)

    [(myl) But hard to VERB NOUNPHRASE doesn't in general mean "it is difficult to conceive of a desirable outcome if NOUNPHRASE were to be VERBED". "It's hard to open this jar" doesn't mean "It is difficult to conceive of a desirable outcome if this jar were to be opened".

    You seem to be relying on the Humpty-Dumpty theory of meaning here:

    "When I use a word," Humpty Dumpty said, in rather a scornful tone, "it means just what I choose it to mean- neither more nor less."

    "The question is," said Alice, "whether you can make words mean so many different things."

    "The question is," said Humpty Dumpty, "which is to be master-that's all."

    Alice was too much puzzled to say anything; so after a minute Humpty Dumpty began again. "They've a temper some of them- particularly verbs: they're the proudest- adjectives you can do anything with, but not verbs- however, I can manage the whole lot of them! Impenetrability! That's what I say!"


  40. Jeroen Mostert said,

    September 18, 2012 @ 8:11 am

    @myl: yes, thank you. You've made me realize that my explanation is insufficient, in that it doesn't really explain anything, but just draws up an alternate interpretation without telling us how we would tell it's valid (which would be an actual explanation). I'm almost considering a study in linguistics to see if this would improve my reasoning. :-)

  41. DEP said,

    September 18, 2012 @ 10:51 am

    This from today's news:
    "I can't underestimate how important the Fusion is to Ford," Fields told workers at Flat Rock last week, noting the segment is growing two times faster than the rest of the industry.


  42. George said,

    September 18, 2012 @ 3:50 pm

    @DEP. Well spotted. Maybe it's just me but I find it sounds wrong in the first person. Something about the idea that I would be telling myself that I can't (as in "must not") underestimate something just doesn't work. So, for me at least, it's just plain wrong.

  43. boynamedsue said,

    September 19, 2012 @ 4:49 am

    Again with the presciptivism.

    For very many speakers of English can not=must not, except in cases of ability, where can must be used.

    'People can not behave like that' has a near equivalent meaning to 'People must not behave like that'.

    'People mustn't underestimate the importance of the famine' = 'People cannot underestimate the importance of the famine'

  44. Ted said,

    September 19, 2012 @ 11:39 am

    As the Great Irish Famine said in 2000, "They misunderestimate me."

  45. Ted said,

    September 19, 2012 @ 12:48 pm

    @Jeroen and myl: I actually think that "Its importance cannot be underestimated" isn't at all equivalent to "Its importance can be underestimated only with difficulty."

    The problem here arises, I think, with the author's use of the passive voice — an instance where, pace Pullum, Strunk and White are onto something. (Prof. Pullum can at least take some small comfort from my accurate description of this construction as passive.)

    I suspect that this is a rhetorical device intended to convey authority, either by mimicking the style of academic writing or by purporting to describe an objective reality independent of the identity, or even existence, of the observer (which, come to think of it, may be the reason this style is characteristic of academic writing). But it has the unfortunate consequence of making the deontic modal interpretation difficult to support, because without a subject who is capable of suffering negative consequences or ethical discomfort, it reads as a simple statement about what is possible.

    I think the intended meaning is something more like "One must guard against the tendency to underestimate its importance," which can be simplified as "One must not underestimate its importance." This, in turn, is equivalent to "One cannot underestimate its importance" (in the deontic modal usage), which is then passivfied into "Its importance cannot be underestimated."

    In your discussion, you've reactivated it as "We cannot underestimate its importance," then analyzed that version by comparison to "It is difficult to underestimate its importance." The latter of course, is the opposite of what is meant, hence the conclusion that there's a misnegation at work here. But the problem may be instead that the analogy to "It is difficult," while tempting, is inapposite and may be leading you astray.

    A better point of departure, ironically, might be "It is easy to underestimate its importance, but one cannot do so," which is oxymoronic on its face but is saved from meaninglessness by the deontic modal interpretation.

RSS feed for comments on this post