WHO: 5 percent of calories should be from sugar

« previous post | next post »

Even though I've been reading that headline on my portal page for 3 days now and know what it's really supposed to be saying, I still can't read it the way they intended. The first sentence of the actual article:

The World Health Organization says your daily sugar intake should be just 5 percent of your total calories — half of what the agency previously recommended, according to new draft guidelines published Wednesday.

Even that sentence doesn't really say they'd be happy with 4 percent, or would previously have been happy with less than 10%. But at least the "just" cancels an otherwise implicit "at least". There's a lot of literature about when numbers are interpreted as "exactly" and when as "at least", and about where exactly those two kinds of interpretations come from. But unless they occur with suitable modifiers or in particular constructions, they are never freely interpreted as "at most". So unless we're supposed to believe that WHO wants everyone to get exactly 5% from sugar, that headline is just wrong, I believe.

No big deal. I just had to say it after three days of suffering in silence.



23 Comments

  1. Ginger Yellow said,

    March 7, 2014 @ 8:16 am

    Maybe this falls under "in particular constructions", but isn't the phrasing "x things could be y" almost always interpreted (or meant to be interpreted) as "at most x things are"?

  2. Eric P Smith said,

    March 7, 2014 @ 10:09 am

    @GingerYellow: I'm not with you on "x things could be y". Can you give an example?

  3. Brett said,

    March 7, 2014 @ 11:08 am

    @Eric P Smith: How about "A thousand people could be infected"?

  4. John Finkbiner said,

    March 7, 2014 @ 11:19 am

    @Eric P Smith: Generally I see the construction in news reporting. I can't figure out how to search for examples so I'll make some up:

    1) Dr Smith said that 1 in 5 Americans could be at risk for DOTD*.

    2) Twenty percent of BigCorp employees could find themselves out of a job if the merger with LargeCo is approved.

    3) One in ten Philadelphians could be wasting money on worthless advice.

    4a) A fire department spokesman said that fifteen people could have been injured in the fire.

    4b) A fire department spokesman said that fifteen people could have been injured if the fire had not been contained so quickly..

    Generally I take this construction to mean "as many as [number] things could be (or have) [attribute]."

  5. Jonathan Mayhew said,

    March 7, 2014 @ 11:40 am

    Most will understand "just' as "no more than." Sugar is bad, the thinking goes, so it's not that you need to get at least 5%. If you said "green vegetables" then the message would be opposite.

    The World Health Organization says your daily vegetable intake should be just 5 percent of your total calories — half of what the agency previously recommended, according to new draft guidelines published Wednesday.

    Here is the meaning is you still need veggies, but not as much. What is unclear is whether more veggies would be harmful, or just not necessary.

  6. Eric P Smith said,

    March 7, 2014 @ 12:33 pm

    @Brett

    I see. I'm not sure though that "A thousand people could be infected" implies that more could not be. It may transpire that the danger of infection is worse than formerly thought, and a later report could conclude "Ten thousand people could be infected" without contradicting the first report.

  7. Ginger Yellow said,

    March 7, 2014 @ 12:52 pm

    I see. I'm not sure though that "A thousand people could be infected" implies that more could not be. It may transpire that the danger of infection is worse than formerly thought, and a later report could conclude "Ten thousand people could be infected" without contradicting the first report.

    It wouldn't contradict it, but it would reflect different information. As a journalistic construction, you wouldn't write "1000 people could be infected" if you thought 10000 people could be. In journalese, at the very least, "could" signifies a maximum (or a plausible maximum, anyway). It's an arse-covering usage to get the boldest possible claim in print, without being technically untrue when the truth turns out to be less dramatic. You see it all the time in the British press. "1 in 10 homes could be flooded in 20 years", that sort of thing..

  8. Jenni said,

    March 7, 2014 @ 12:53 pm

    Whenever the World Health Organisation weighs in on anything I instinctively assume Pete Townshend has started doling out health tips

  9. John Lawler said,

    March 7, 2014 @ 12:54 pm

    While it was hardly a model of comprehensibility, I had no trouble parsing the sentence this way:

    your daily sugar intake (calories)Q[sugar]
    should be just
    5% of your total (intake) caloriesQ[total]/20
    I.e, Q[sugar] ≤ Q[total]/20

    It's the just in should be just that provides the inequality . It's a vector quantifier, like the difference between few people and a few people.

  10. Barbara Partee said,

    March 7, 2014 @ 1:07 pm

    John, sure, the sentence in the article with the "just" in it was fine — my problem was with the headline, in which they omitted the "just". I think you need something additional, such as the "just", to get an "at most" reading.

  11. Barbara Partee said,

    March 7, 2014 @ 1:19 pm

    @ Ginger Yellow, yes, under "particular constructions" I did have in mind some constructions with modals, like "Can you live on $20 per day?", "He can run a mile in 4 minutes". (Consider being offered a prize if you run a mile in 4 minutes. You won't get the prize if you run it in 4.2, but you will if you run it in 3.8.) I hadn't thought of "1000 people could be infected", but yes, as you note, something similar happens there too. That's an interesting one, because it's not paraphrasable with either "at least 1000" or "at most 1000". A common locution is "As many as 1000" — and if you wanted to indicate that it's a small number you could write "As few as 1000" (as in "By 2020 there could be 5 speakers of that language left".) Those are interesting because they indicate what John Lawler was calling the 'vector' (I think), but they don't actually say "at least n" or "at most n" — I think they really say that it's possible that there will be n, and then give an idea of whether other likely possibilities involve bigger or smaller numbers. I don't even know if that has been studied.

  12. KathrynM said,

    March 7, 2014 @ 1:24 pm

    I certainly read the headline as "you need to consume enough sugar each day so that 5% of your caloric intake comes from that sugar." Which made both my ears and my eyebrows perk up!

  13. Jerry Friedman said,

    March 7, 2014 @ 1:46 pm

    I agree that the headline is misleading, but I wonder whether it's based on some assumption that "obviously" "everybody" gets more than 5% of their calories from sugar and wants as much as possible, so "You should get no more than 5% from sugar" can be paraphrased as "You should reduce your sugar to 5%."

  14. great unknown said,

    March 7, 2014 @ 2:22 pm

    This is used deliberately in advertising:

    "Save up to 50% or your money back!"

  15. GeorgeW said,

    March 7, 2014 @ 2:52 pm

    I suspect that this was written by someone who is not a native English speaker but well enough versed in English to write this type material (in English).

  16. Azimuth said,

    March 7, 2014 @ 3:19 pm

    "WHO: 5 percent of calories should be from sugar"

    could have been:

    "WHO: Max 5% calories should be from sugar"

    The "max" is accurate and snappy. Who needs "of"? Let's bring "percent" back to its roots as a per-hundred counter. (I tolerate "half of" or "much of". "Percent of" sounds barbaric if you think about it too much.)

  17. C said,

    March 7, 2014 @ 3:27 pm

    I certainly read the headling as implying my calorie intake better contain at least 5% sugar or else. Basically along the lines of "you should drink X glasses of water a day" — the implication being that I'm probably not drinking X glasses of water a day and I'd better start to.

    I would count this to be an extraordinarily poorly crafted headline given the actual intent of the author.

  18. Milan said,

    March 7, 2014 @ 3:34 pm

    @Ginger Yellow:
    I think, X in a construction like "X could be Y" is interpreted as an exact number, but in practice it is a maximum, because "could" denotes only a possibility. It basically says: "There are X [entities] that are possibly Y (but also possibly not-Y)" Clearly, there do not need to be X entities that actually are Y to make this statement true.

  19. BobC said,

    March 7, 2014 @ 4:19 pm

    My problem is in trying not to interpret the "just" in "just 5%" as "merely" or "only," as in "there were just 5 people ahead of me on line."

  20. Jason Merchant said,

    March 7, 2014 @ 4:30 pm

    As Jerry Sadock once pointed out to me, the conventions of scoring in golf lead to default "at most" interpretations: "If you get an 85 on this course, you will qualify for our tournament" allows for qualification with scores less than 85, but not more than 85. (As opposed to the more usual lower-bounded reading in e.g. "If you get an 85 on this test, you will pass this class".) Sugar=golf?

  21. Jonathan Mayhew said,

    March 7, 2014 @ 5:21 pm

    Right, so you have to be twelve to order the kid's meal that means not more than 12. You can't call a forty-minute mile a 4-minute one, on the logic that you've beaten four minutes by 36 extra ones. To someone who's just shot 75 in golf, you can't say, "That's nothing, I shot 120"

  22. Brian Buccola said,

    March 9, 2014 @ 9:10 am

    @Barbara Partee: It's interesting that you write, "Even that sentence doesn't really say they'd be happy with 4 percent." Indeed, normally, I think, "should…just n" means "should…n and no more than n", i.e. "should…exactly n", rather than simply "should…no more than n":

    (1) You should buy just five apples.

    Suppose John is going to the store, and Mary tells him (1). If John comes back with six (or more) apples, then he certainly didn't do what Mary instructed. But even if he comes back with four, three, … or certainly zero apples, then I think he likewise disobeyed Mary's instruction. It's possible, however, that this lower bound, unlike the upper bound contributed by "just", is an implicature. Compare (2) with (3).

    (2) You should buy just five apples. #More is okay, too. — feels contradictory

    (3) You should buy just five apples. Less is okay, too. — does not feel as contradictory

    Curiously, I think "should" works differently from "have to" in this respect:

    (4) To win the game, you have to touch just five rocks.

    For example, if it's a game where you need to get from point A to point B by jumping from rock to rock, in as few jumps as possible, then (4) means "five or less" (similar to the golf example); and so a follow-up like "Less is okay, too" works fine. But if it's a game where each rock you touch gives you one point, and you need (at least) five points to win, then (4) means "five or more", in which case a follow-up like "More is okay, too" works fine.

    Lastly, there has been some discussion of lower bound numeral modifiers, e.g. Nouwen 2010. In a reply paper to Nouwen, my co-authors and I discussed some contrasts between "at most n" and "up to n", including a brief footnote on examples like "Save up to 50%!" vs. "??Save at most 50%!" (which was mentioned in a comment above by great unknown). One thing I find interesting is that English has no modifier expression "down to". So for example, (5) could describe, say, the positive future prospects of some invented language, while (6) can be used to describe the negative prospects of a dying language, but not with "down to".

    (5) By 2020, there could be {as many as / up to} 100 speakers.

    (6) By 2020, there could be {as few as / *down to} 5 speakers.

  23. Barbara Partee said,

    March 9, 2014 @ 12:19 pm

    @Brian – Nice, thanks. Yes, that's the sort of work I had in mind when I wrote "There's a lot of literature about when numbers are interpreted as "exactly" and when as "at least", and about where exactly those two kinds of interpretations come from." And of course I should also have included
    'and when as "at most".'

RSS feed for comments on this post