« previous post | next post »

According to a press release posted yesterday by Michael Eisen on the NY Giants web site, the team has released Plaxico Burress, the wide receiver who famously shot himself in the leg at a nightclub. Eisen cites the player's achievements:

Burress is perhaps best known for catching the game-winning touchdown pass in the Giants’ upset victory over the New England Patriots in Super Bowl XLII. He eluded Patriots cornerback Ellis Hobbs with an inside move, then ran to the outside, where he caught Eli Manning’s 13-yard throw with 35 seconds remaining in the game. Burress had two receptions in the Super Bowl, catching Manning’s first and last passes of the game.

In his four years with the Giants Burress caught 344 passes, which places him 12th on the franchise’s career list, one catch ahead of Earnest Gray and three behind Aaron Thomas. Burress had 3,681 receiving yards and caught 33 touchdown passes for the Giants.

And then he quotes the coach:

“Plaxico’s contribution to our championship season in 2007 can never be underestimated or undervalued,” said Head Coach Tom Coughlin.

This is a fine example of an issue that we've discussed many times — see here for a summary and some links. And Plaxico is not the only person in recent news reports and press releases whose contributions cannot, can't or can never be underestimated. At the 2009 Eaton Science Fiction Conference, for example,

"We are really delighted to be able to host an Eaton conference with this emphasis on Jules Verne,” said Melissa Conway, head of Special Collections and Archives at UCR. “Verne’s influence – right up to the present day – cannot be underestimated."

And a discussion of Michael Jackson's comeback tour in the U.K. tells us that "his drawing power can’t be underestimated". I knew about Jules Verne and Michael Jackson, but not about Cindy Sherman, though Alexis Loinaz tells me here that "Cindy's prowess just can't be underestimated"

In fact, the sum of Google News hits for "cannot be underestimated" (219) "can't be underestimated" (112) and "can never be underestimated" (256) is 587, and without checking them all, I'll be that they all mean "must not be underestimated" or "cannot be overestimated". This compares to 40 for the logically correct "cannot be overestimated" (29) "can't be overestimated" (9) and "can never be overestimated" (2); or 18 for "must not be underestimated" (17) "mustn't be underestimated" (1) and "must never be underestimated" (0). The only logically sound alternative that's competitive is "should not be underestimated" (618) "shouldn't be underestimated" (54) and "should never be underestimated" (46), which wins out in this morning's count by a mere 718 to 587.

Does this mean that "cannot be underestimated" should be considered an idiom, like "still unpacked", "fail to miss", and "could care less"? I don't think so, at least not exactly. One problem is that the same pattern applies to "impossible to underestimate", "hard to underestimate", "difficult to underestimate"; to "cannot understate", "cannot undervalue", "cannot underrate"; and so on.  In the other direction, expressions like "should not be overestimated" are also often used to say the opposite what they literally mean. Whenever we combine negation, concepts of possibility or difficulty, and thresholds on a scale of evaluation, people seem to get their wires crossed in the same way.

Some people argue that all these examples are logically correct – that is, they actually do mean what their authors intend them to mean —  if you only analyze their semantics in the right way. I incline towards the "poor monkey brains" theory myself.

[Please read this, at least, before commenting.]


  1. Campbell said,

    April 4, 2009 @ 2:13 pm

    You suggest "must not be underestimated" as something which might be meant by "cannot be underestimated". But this seems a tautology. To underestimate something is to estimate it more lowly than it should be estimated. So we should not, or must not, underestimate anything.

  2. Ellen K. said,

    April 4, 2009 @ 5:04 pm

    Campbell, sometimes, underestimating someone isn't a problem. We don't need to be perfect at everything all the time. We don't have to be perfect at estimating. There are times, though, when it's important not to underestimate.

    Times when it's better not to underestimate, but no big consequences if we do. And other times when we must not underestimate, because the consequences would be bad.

  3. Campbell said,

    April 4, 2009 @ 11:32 pm

    Ellen, good point. But I doubt that "x cannot be underestimated" often means that underestimating x will have bad consequences, as you suggest. When people say "x cannot be underestimated", they usually mean to convey that x is especially good or impressive in some way. But in general there's no reason to think underestimating good things will have worse consequences than underestimating mediocre things.

  4. Noetica said,

    April 5, 2009 @ 8:17 am

    I am reminded of no question that P (NQT) versus no question but that P (NQBT), when the meaning is "no doubt that P". Here too it's unclear whether, where, or how negation of P is introduced.

    In Australia it is hard to avoid tracking the evolution of American political rhetoric, and our own politicians' mimicry of it. Not only do American leaders incorrigibly "stand shoulder to shoulder with x", or "stand ready to Φ" – ours have begun to do the same. Not only do the American people "have no greater friend than x" (a neat and diplomatic way of having several best friends in the world) – so, perforce, do we.

    Now, I have observed that NQBT was dominant up till about ten years ago, but that it is now almost totally replaced by NQT in American usage as filtered through the media. The wonderful M–W Concise Dictionary of English Usage notes these two forms, but not the evolution I have remarked upon. It takes both to be adequate for expressing "no doubt that". Garner agrees that they both clearly mean the same, and he therefore calls for omitting the but as redundant (A Dictionary of Modern Legal Usage, p. 597).
    Myself, I'm with the humble Telegraph Style Book on this one:

    question: no question but that = No doubt that. No question of = No possibility of. It is correct but ambiguous to write "no question that" or "no question but". Best avoid.

  5. Ellen K. said,

    April 5, 2009 @ 12:28 pm

    @Campbell: No, I did not suggest that "x cannot be underestimated" often means that underestimating x will have bad consequences. Or even that it ever means that. I was commenting on the idea that "must not underestimate" is a tautology.

  6. Jan Schreuder said,

    April 5, 2009 @ 2:54 pm

    I agree with Mark that cannot underestimate usually is intended to mean cannot too highly value and thought the same about Coughlin's remark. However, after reading what Mark wrote in the the posting he refers to at the end of today's contribution, I start to hesitate about the Coughlin citation. Because the Burress affair bristles with moral implications and has been the occasion of a lot of preaching and moral grandstanding. Coughlin might have wanted to imply that people would like to forget Plaxico's record because of his off field foolishness, but they should not underestimate his contribution to the Giants' championship. Thus cannot be underestimated might have been intended to mean should not be underestimated

  7. Simon Cauchi said,

    April 6, 2009 @ 2:46 am

    Here's another example of someone saying the opposite of what they really mean. The price of milk from the smaller dairy companies in New Zealand is (at the moment) cheaper than milk from the big amalgamated firm Fonterra. Why? "Because we don't have Fonterra's economies of scale."

  8. bianca steele said,

    April 6, 2009 @ 9:59 am

    we seem to get our wires crossed in the same way

    No doubt, though I'm not sure that's what's happening here.

    Most boringly, after looking at these posts, I'm not sure I buy the "traditional" can/may distinction. I think my usage is pretty good, and I do use "can" at times. (For example, I don't ask my husband, "may I use the bathroom," when what I mean is, "have you already shaved?" I say "can" in that case. When I'm being careful, I do say, "may I have that knife after you're done with it?")

    I do think there is confusion as to whose thoughts are being discussed. This shows up well in the quotation from Alessandra Stanley in one of the older posts. She is not talking about her readers' need not to underestimate the writers' task, she is talking about the writers' need not to underestimate their task. She is talking, in other words, from the writers' viewpoint, about their work — she is not, in that sentence, primarily articulating what readers should know. (The assumption here is that readers of criticism aren't interested in the psychology of the artist or the process by which the art was made.) Several of these examples are of people who may be the ones whose behavior is being evaluated, or may be evaluating the behavior of others — it would depend on the structure of their organization — using language that, thus, might be more appropriate to someone with a different relationship to those people than the one they have.

    Also, you will occasionally read someone saying that "I cannot do this or that" indicates an unwillingness to accept responsibility. A professor might complain that a student says, "I cannot make it to class next week." He would prefer to hear, "I will not make it to class next week," indicating a choice on the student's part, that the student recognizes he or she has a will of his/her own. Others may have different preferences, however.

  9. Språkets betydelse kan inte underskattas said,

    April 11, 2009 @ 8:16 am

    […] Låt mig också säga att detta inlägg fick skjuts av att Language Log nyligen tog upp uttrycket utifrån dess (påstådda) ologiskhet – här. […]

RSS feed for comments on this post