Stupid less/fewer automatism at the WSJ

« previous post | next post »

Spot the horrible effect introduced here by an over-picky Wall Street Journal subeditor:

Quite often, these games don't even turn out to be good: Fewer than half of them have been decided by 10 points or fewer.

That "10 points or fewer" phrase on the end is a desperate and quite ridiculous effort at obeying the prescriptive rule that you should use fewer for all things that can be counted, and less only for mass quantities.

I should stress, to forestall misunderstanding, that I agree there is a difference in the distribution of less and fewer. I myself would say You should drink fewer beers and You should drink less beer. And certainly I find *You should drink fewer beer unthinkable. But there is an asymmetry here, which is one of the many things the prescriptivists fail to realize: while *fewer beer is completely ungrammatical (for every American speaker, I would think), less beers is merely informal. A guy on a 3-day week at the construction site who says I'll just have to buy less beers on a Friday night isn't showing that he can't tell less from fewer; he knows the difference perfectly well (when the boss says he'll be working fewer hours he knows exactly what that means), but the range of cases where he will choose less is wider than it is for some other speakers (a bit wider than the range I would accept, in fact).

However, with many countable objects it just isn't true that less is unacceptable for most Standard English speakers and writers, even for educated speakers using formal style. Take counting of time periods in days, for example. Here the prescriptive rule is particularly far away from matching normal usage: Your package will arrive in seven days or less is far better than ?Your package will arrive in seven days or fewer; but the prescriptivists and subeditors seem deaf to such matters. They want an easy mechanical rule. (Language Log first treated this topic at length here; see also this post and this one, about supermarkets intimidated into changing their "10 ITEMS OR LESS" signs.)

In the Wall Street Journal example above, the quantity referred to is a points range, and less would have been ideal: everyone accepts less than 11 points, and there is nothing wrong with 10 points or less in this context either.

What makes the miscorrected version particularly risible is that the miscorrection creates an echo: the clause after the colon both begins and ends with fewer, and sounds appalling as a result. But the subeditor was deaf to this fact too.

It's not the first time, by any means, that I have seen something grossly unacceptable published because someone felt that the more acceptable version violated some kind of misbegotten prescriptive rule out of Strunk & White or some other vile manual of bad advice, or some myopic subeditor changed something without even reading the whole clause — without looking at even a dozen preceding words of context.

I almost despair when I see this sort of asininity. America seems to be a nation of cowed and abused users of the English language, constantly fearing punishment for violation of crassly inaccurate and simplistic rules that they do not understand, rules enforced by stupid style manuals, ignorant writing instructors, dimwitted editors, and their cowardly henchmen.

Rise up, America! You threw off the chains of British rule in 1776! Throw off the chains of the prescriptive poppycock that originates from the very same country in the very same century!

Oh… what's the use. A whole lot of you are probably going to write in the comments area below that you actually like the Journal sentence quoted above, and approve of the mechanical less/fewer rule. I give up.

Hat tip to a brave and rebellious American and WSJ reader, Ron Irving, of Seattle.

Share:



116 Comments »

  1. Nathan said,

    December 2, 2010 @ 6:12 pm

    "10 points or fewer" is horribly, unacceptably bad English.

  2. Alex Poterack said,

    December 2, 2010 @ 6:29 pm

    I perfectly understand your point and your frustration, but I must admit; when I learned the rule for distinguishing less and fewer, that is, the simple mechanistic one, I took to it immediately, and anything that doesn't follow it sounds wrong to me, the same way, say, dropping an article at the wrong time sounds wrong. Strangely, this only started after I learned the rule; I was perfectly blissful in my prior ignorance. So I'm perfectly okay with "10 points or fewer", although I see your point about the echo, and"10 points or less" sounds wrong, and isn't something I'd say (a similar and common example that rankles me is when baseball announcers refer to there being "less than 3 outs", which never fails to annoy me. Again, I know it's silly presciptivism, but it's become ingrained for me; my reaction is immediate and unconscious.

  3. Megs said,

    December 2, 2010 @ 6:29 pm

    Actually I find the entire construction of "or fewer" to be a bit of an anachronism. Fewer just doesn't sound right on the ear on the end of a phrase and it's horrendous in sports, grocery stores, or duplicate. :shudders: (And I'm like you in how strong I am about following the rule in general.)

  4. Iulus said,

    December 2, 2010 @ 6:47 pm

    I don't really see the motivation for reviling "or fewer" except as emotional backlash against prescriptivism. "Or less" is certainly more common, but both are perfectly grammatical here.

  5. Gerri said,

    December 2, 2010 @ 6:49 pm

    You may beat your head against the wall of die-hard prescriptivists, but don't give up. There are plenty of American wordsmiths who agree with you. They're the same ones who don't despair when seeing a preposition at the end of a perfectly good sentence. I suspect most call themselves copy editors, not subeditors, but that's picking nits from across the ocean.

  6. Mark Flowers said,

    December 2, 2010 @ 6:52 pm

    @Iulus – I would almost agree with you, except for the echo that Prof. Pullum noted. I didn't hadn't even fully parsed the sentence to see which fewer was misplaced before I cringed at the wrongness of the double "fewer."

  7. PaulAtNorthGare said,

    December 2, 2010 @ 7:00 pm

    A couple of random thoughts, from someone for whom the count/mass (or discrete/continuous) distinction between the usage of "fewer" and "less" just seems to work well in almost all cases.

    First, I wonder to what extent the examined sentence jars primarily because of the repetition of "fewer", and less because of the second use of "fewer". For example, this would actually read okay to me:

    "Less than half of them have been decided by 10 points or fewer."

    I'd actually argue that "Less than half…" works there because "half" – quite apart from what it's measuring in context – is a point on the *continuous* scale that fractions represent (it would be hard to argue that "a third" – 0.3 recurring – *isn't* on a continuum). Without the echo, the later use of "fewer" doesn't jar for me, and the above then follows the count/mass heuristic very well.

    A second point is that I wonder if the acceptability of "ten points or less" is helped by the fact that *points* (like time) are an abstract concept, not bound to the existence or otherwise of physical objects. Notwithstanding the rules of any particular sport, it's perfectly okay to have fractional points (think of Olympic figure skating or diving), so in the abstract "points" are on some level a continuous measure, so we're happier to see the word used with "less" than we are real, concrete objects.

    Not trying to be prescriptivist here. I just think that the count/mass heuristic works extremely well in most cases, and with some careful thought more cases than are immediately apparent.

  8. Xmun said,

    December 2, 2010 @ 7:07 pm

    How about these two examples? They don't set off any "that's not right" alarm bells for me:

    http://www.gnn.com/article/on-stage-in-140-characters-or-fewer/983191

    “This work is in the public domain in the United States, and those countries with a copyright term of life of the author plus 100 years or fewer.”

  9. KCinDC said,

    December 2, 2010 @ 7:10 pm

    I find "10 points or fewer" a lot less ungrammatical than, say, "He makes fewer than $30,000" (or "She's fewer than 5 feet tall", though I've never actually encountered anything like that one), where we're clearly talking about an amount or measure. Points seem to have a little more individual existence than dollars or feet, and are more indivisible.

  10. Mark P said,

    December 2, 2010 @ 7:14 pm

    I'm not a sports fan, but I suspect there might be a reasonably well-followed practice in this sort of thing. I can imagine a hockey commentator saying that one team scored fewer goals than another. But I can't quite imagine an American football commentator saying that one team scored fewer points than another. Along the same lines, I can imagine a football commentator saying that a team scored fewer touchdowns, but not less touchdowns. It seems have something to do with the magnitude of the number in question. Hockey scores are usually low, while football scores are often high. So, larger numbers use less and smaller numbers use fewer?

  11. James said,

    December 2, 2010 @ 7:17 pm

    My hunch is that "less" sounds perfectly fine with a count when it is natural to take it to modify the measurable property that's measured with a unit and a number. For instance, it sounds completely fine (to me) to say, "She'll be here in twenty minutes or less," and my speculation is that it's because it means, "… or less time." (Or course, I would not actually say, "She'll be here in twenty minutes or less time," but if someone asked, "Twenty minutes or less what?" I would answer happily, "time".) Notice, in support of this hypothesis, that it's unquestionably fine to say, "She will be here in an hour or less," or even (though not quite as natural) "She will be here in one hour or less." (Just try it with "fewer" — impossible!)

    I'm not exactly sure what points measure. Height of the characters, I suppose. So maybe that's why "10 points or less" is unquestionably fine.

    Of course, my hypothesis doesn't explain "Twelve items or fewer" (which I agree also sounds fine), but maybe that's more of a fixed phrase or idiom?

  12. maidhc said,

    December 2, 2010 @ 7:20 pm

    The sentence would sound better using a different construct such as "within a 10-point margin" (off the top of my head) for the second one. The repetition not only sounds awkward, it makes the sentence harder to read.

  13. Dan K said,

    December 2, 2010 @ 7:27 pm

    To my ear, "your package will arrive in 7 days or fewer" is not only less idiomatic and less comfortable, it also conveys, incorrectly, that your pacakge will arrive in an integral number of days. Perhaps a refund of shipping fees would be in order when such a promise is broken.

  14. Ellen K. said,

    December 2, 2010 @ 7:29 pm

    I'm with Megs on "or fewer" sounding odd.

    And, Iulus, no one is reviling "or fewer". No reviling from me, nor do I see that in Megs comment. And she was the only one (before your post) who commented on "or fewer" in general, rather than a specific phrase with it.

  15. Russell said,

    December 2, 2010 @ 7:38 pm

    Random observation: a great deal less (X) sounds reasonably OK, but a great deal fewer (X) sounds much worse. I wonder why. (COCA counts: 22 for less, 1 for fewer)

    (also fun: a great many fewer, ?a great many less)

  16. Faldone said,

    December 2, 2010 @ 8:10 pm

    What I've never had any prescriptivist be able to answer is why we need less for mass nouns and fewer for count nouns but more works for both.

  17. GeorgeW said,

    December 2, 2010 @ 8:40 pm

    It is my impression that 'fewer' is dying a slow death. It seems that generally when there is a violation of the count/mass rule, it is with 'less' rather than 'fewer' (When in doubt, default to 'less'). This may be why the 'fewer' in the WSJ material is so jarring. And, it is reasonable, as has been suggested, to suspect that a mechanical application of prescriptivism is the reason.

  18. Tom Roland said,

    December 2, 2010 @ 9:15 pm

    Trader Joe's is prescriptivist extrardinaire. http://www.seriouseats.com/2010/01/trader-joes-10-items-or-fewer-sign.html

  19. Garrett Wollman said,

    December 2, 2010 @ 9:42 pm

    I scrupulously observe the distinction in formal writing, but in most situations I neither notice nor care. "Or less" is close enough to being a fixed phrase that I would accept in places where I would not accept "less than".

  20. Coleman Glenn said,

    December 2, 2010 @ 10:22 pm

    I think James has it exactly right, and I think it does apply even in less obvious situations than something like time. If there's a sign that says "12 items or less" I'm not thinking "12 items or fewer items" – I'm thinking "12 items or a smaller quantity." Maybe that's one of the reasons for a shift toward "less": using "fewer" makes you subconsciously restate the subject – a split second of extra work – whereas less refers to a more general concept of "not as much." The mind takes the path of least resistance. I think that has something to why I tend to prefer "less," anyway.

  21. Nathan Myers said,

    December 2, 2010 @ 10:39 pm

    For my part, the examples that are clearly wrong are the more interesting. "Fewer beer" indeed.

  22. Jon Lennox said,

    December 2, 2010 @ 10:53 pm

    GeorgeW: "why we need" is rarely a meaningful question for natural language. There's no particular reason why English "needs" the distinction between "much" and "many", but it has it.

    Interestingly, despite my usual (prescriptivist) rule of thumb that "less" / "fewer" reflects "much" / "many", I agree that the sentence "Boston didn't score many points, but New York scored less" is still reasonably well-formed. (Though "…New York scored fewer" would also be well-formed.)

  23. Lindsay R. said,

    December 3, 2010 @ 12:22 am

    it seems to me that the issue is less about the semantic role of fewer vs less than it is about syntactical elements winning out over connotation. The cases cited where less is preferred by speakers but fewer more closely adheres to the 'rule' are ones in which there is a countable noun antecedent but the implication of less is of an unknown or continuum encompassing amount (see JamesOfNorthGate's comment).
    Compare the ubiquitous '10 items or less' which means and number of items under 10 with the game example about 'Boston scored few points but New York scored even fewer:' in the 10 items of less scenario the number of items is unknown and can vary. In contrast the fewer points scored by New York cannot vary. It is a fixed number that simply isn't stated.

    Could it be the brain distinguishing between fixed and non-fixed quantities – kinda like perfect/imperfect tense – rather than simply taking the path of least resistance?

  24. Andrew Shields said,

    December 3, 2010 @ 12:27 am

    "But there is an asymmetry here":

    The asymmetry comes up with "number" and "amount" as well: "A number of people" is common, but "an amount of people" is pretty frequent as well, while "an amount of knowledge" is common, but "a number of knowledge" is nonsense.

    (I tried to back this up with some loose Google-hit data, but "a number of knowledge" generates a lot of hits with phrases where "knowledge" modifies a noun that follows it.)

    Has there been any further discussion on Language Log of such asymmetries?

  25. Dan K said,

    December 3, 2010 @ 12:39 am

    Re: amount/number, I wonder if this is related to how people will convert countable things to mass things by suffixing with -age (sorry if that's not the right terminology). As in, when you arrive at a stadium, "there's a large amount of peopleage here." Of course the word "peopleage" doesn't sound pleasant to anyone, but it's quite clear you can't talk about the number of peopleage.

  26. Mark Sirota said,

    December 3, 2010 @ 12:59 am

    It may be that "less than 11 points" sounds okay because we're familiar with the mathematical comparison operator "less than" (<). So "Ten items or less" in the checkout line sounds wrong, but "less than 11 items" is almost okay (the "items" unit makes it countable, so it sounds okay until you get there.)

  27. Trevor Barrie said,

    December 3, 2010 @ 1:10 am

    Mark P: Well, that's because the hockey commentators are likely Canadian, and hence better educated than their American counterparts.:)

    More seriously, I think PaulAtNorthGare has it. It's not whether the noun is grammatically a mass noun or count noun that matters, it's whether the speaker conceives of the number as representing a count of discrete objects or a measurable quantity. "Goals" and "touchdowns" are clearly discrete events; 3.2 touchdowns is simply meaningless. But "points" are an abstract quantity; a specific game may not make it possible to score 3.2 points, but the concept is certainly meaningful.

    This explanation also explains why you can probably get away with using "less" anywhere, but there are cases where "fewer" is clearly wrong. Any time you're counting something, higher numbers still denote more "stuff": fewer touchdowns means less touchdownage, to use Dan K's terminology. You can't always make the reverse substitution: if you're talking about height, duration, or mass, you're clearly measuring and not counting, so "fewer" won't do.

    Faldone: You make a good point. I may have to start distinguishing between "more" and "manier".:)

  28. Xmun said,

    December 3, 2010 @ 1:10 am

    No need for "there's a large amount of peopleage here." Just say "Gee, what a crowd!"

  29. C Thornett said,

    December 3, 2010 @ 1:37 am

    'Or fewer' does seem to be the problem here. (Look, Ma, no grammar mistakes!) Do others find that '10 or fewer points' sounds less clunky than '10 points or fewer'?

    Perhaps English uses 'less' as a kind of default in certain circumstances which might include final positions and cases where countable/uncountable is not specified or is not clear.

  30. Pflaumbaum said,

    December 3, 2010 @ 2:17 am

    Jon Lennox said:

    GeorgeW: "why we need" is rarely a meaningful question for natural language. There's no particular reason why English "needs" the distinction between "much" and "many", but it has it.

    It was Faldone you were quoting btw, not George W. I think the underlying point was that prescriptivists police their wide-ranging less/fewer distinction on the grounds that letting it lapse would make the language confusing; but they can't answer the question of why no such confusion arises with the comparative, for which we use a single word. (Or, incidentally, how other languages make do perfectly well with a single word for the positive).

  31. Peter Taylor said,

    December 3, 2010 @ 2:25 am

    As a Brit, am I allowed to prefer "10 points or fewer"? Is it only Americans who would make you despair by expressing such an opinion?

    (For what it's worth, here I would favour "no more than 10 points").

  32. Pflaumbaum said,

    December 3, 2010 @ 2:27 am

    Sorry, got my terminology jumbled there (it's early and I'm out of coffee). They're both comparatives obviously. I meant 'why no such confusion arises with its opposite'. Er, I think…

  33. Pflaumbaum said,

    December 3, 2010 @ 2:40 am

    @ Peter Taylor

    But doesn't the fact that you would have to recast the sentence suggest that you're uncomfortable with fewer here?

    For some reason less/fewer prescriptivism never got through to me, even though in other respects I was quite a little pedant growing up. I made an effort not to say things like Me and Mike are going to the match because I could see the (supposed) grammatical logic, but wide-ranging less/fewer just made no sense and seemed totally artificial.

    I'm British too btw (North London, RP)

  34. Barrie England said,

    December 3, 2010 @ 2:44 am

    Pam Peters (‘Cambridge Guide to English Usage’, p. 205) writes perceptively on the matter, as on much else: ‘ “Fewer” draws attention to itself, whereas “less” shifts the attention on to its more significant neighbours.’

  35. George said,

    December 3, 2010 @ 3:18 am

    I think that Dan K touches on something very interesting in his comment of 7.29 pm on 2 December (you now know what side of the Atlantic I'm on) and I'm a bit surprised that nobody else seems to have taken it up. "Vehicles using this bridge must weigh fewer than 10 tons" would imply (to me anyway) that each vehicle must weigh an integral number of tons, from 1 to 10. "Vehicles using this bridge must weigh less than 10 tons" is obviously better, not least because I'm a cyclist. Likewise for "10 tons or fewer" vs. "10 tons or less". But "10 points or fewer" is different – because games are decided by integral numbers of points – and is what I would write myself. If the specific sentence cited is very clunky, it's the repetition of 'fewer' that causes the problem, rather than either of its uses in isolation.

    Both "less than half of the games" and "10 points or less" sound wrong to me, not objectionable-wrong but certainly wouldn't-do-it-myself-wrong. I'd just rewrite the whole thing to avoid the problem.

  36. John Walden said,

    December 3, 2010 @ 3:21 am

    I think I visualise some of these numbers as horizontal lines, with numbers assigned to them, not as real numbers: nobody (nobody?) would say "Water freezes at 32 degrees F, or fewer" and certainly not "Water freezes at at 0 degrees C, or fewer".

    One could argue for "lower" but "less" doesn't raise my hackles here, and google hits agree with me (for what that is worth).

  37. Chris said,

    December 3, 2010 @ 3:45 am

    The WSJ quote certainly sounds clunky to me, because of the inelegant repetition of the word fewer. But what strikes me as really odd is that, in the name of castigating prescriptivism, Professor Pullum should describe the WSJ's preferred style as "grossly unacceptable". If that isn't prescriptivist, what is?

  38. Leonardo Boiko said,

    December 3, 2010 @ 4:01 am

    @Chris: Prescriptivism is like saying you should wear a tuxedo everywhere, even in the beach. Pollum is saying a tuxedo in the beach is unnaceptable. You can call that prescriptivism if you want, but it’s a different kind of prescriptivism; one preaches one single “correct” English everywhere, and the other says to use whatever English is more natural and adequate to the situation at hand.

  39. JohnW said,

    December 3, 2010 @ 4:46 am

    Well, I don't think I've used "fewer" in my life – ever ever. Sounds precious to me. Not that "less" is always the desirable alternative; it's more that there's often a way of rephrasing that circumvents the supposed dilemma. For example, I'd probably have the construction worker say, "A few less beers". And I'd rephrase the WSJ with, "Under half… ten points or less." (British, 40s, RP, me.)

  40. PaulAtNorthGare said,

    December 3, 2010 @ 5:05 am

    Something was nagging away at me here, and I've put my finger on it. This section of Prof. Pullum's piece strikes me as odd:

    Take counting of time periods in days, for example. Here the prescriptive rule is particularly far away from matching normal usage: Your package will arrive in seven days or less is far better than Your package will arrive in seven days or fewer; but the prescriptivists and subeditors seem deaf to such matters.

    I'm a bit baffled by the claim that "the prescriptive rule" would prefer "Your package will arrive in seven days or fewer". If that's true, I completely misunderstand what "the prescriptive rule" is here. My assumption has been that what's typically prescribed is nothing more than the count/mass (or discrete/continuous) distinction, where count/discrete="fewer" and mass/continuous="less". But the corollary of that assumption would be to prescribe "Your package will arrive in seven days or less", because time is a mass/continuous quantity. Never mind that "days" can be counted; they're evenly-spaced marks on a continuum, so "less" ought to be prescribed.

    If the typical prescription is actually that it should be "seven days or fewer", because days are countable, then I'd suggest that the problem here isn't prescriptivism per se, but the fact that the specific prescription is completely clueless mathematically. I've encountered before a certain lack of facility with the abstract concepts of discrete and continuous quantities, and I think this is actually quite significant here. It's one thing to have a nice, simple heuristic that count/discrete quantities prefer "fewer", and mass/continuous quantities prefer "less", but that gets complicated and obscured if it's not clear to people which is which.

    Items at a checkout are pretty uncontroversially count/discrete. Time (& distance, temperature, etc.) ought to be pretty uncontroversially mass/continuous, but it would seem that people's perceptions of whether units of time are discrete or continuous vary quite a lot, and my suspicion is that this derives from huge variance in facility with abstract mathematical concepts. To apply a heuristic – whether prescriptivist or not – to decide whether to use "less" or "fewer", based on the discrete/continuous distinction, one must first be able to make that distinction accurately. I see in the comments above some naive fumbling towards the distinction, and the quote from Pullum suggests that the WSJ folks might have a problem too. The question then is whether the problem is the prescription itself, or their mathematically-hamfisted bungling of it.

    Basically, my hunch is that buried under all of the regional differences and prescriptivism and such, is the fact that in this case the language used is based to some non-trivial degree on ability with mathematics – specifically the ability to grasp the concepts of discrete and continuous measures. A naive approach sees "days" as discrete, countable units of time. A more sophisticated approach (I'd say a more mathematically correct approach) sees "days" as equally-spaced steps along a continuum. Crucial differences in language choice follow.

  41. Tom Saylor said,

    December 3, 2010 @ 5:08 am

    Them’s fightin’ words, Chris. For many people here “prescriptivist” has become little more than a term of derision, reserved for the usage advice of benighted nonlinguists. It can’t possibly apply to the prescriptions of a great anti-prescriptivist like Professor Pullum.

  42. George said,

    December 3, 2010 @ 5:28 am

    What PaulAtNorthGare said. And I hope I'm not one of his 'naïve fumblers'. I like to think that I got most of naïve fumbling out of the way during my teens.

  43. GeorgeW said,

    December 3, 2010 @ 6:49 am

    @Chris: "Professor Pullum should describe the WSJ's preferred style as "grossly unacceptable". If that isn't prescriptivist, what is?"

    As this is a frequent issue on the LL, perhaps a good definition should be posted. I think of prescriptivism as imposing an artificial rule or archaism that is not representative of the speech of a speech community.

    This is particularly egregious when the rule is misapplied as in the WSJ example at hand.

  44. Lou Hevly said,

    December 3, 2010 @ 8:52 am

    There are only 3 instances of "less" in a search at COCA for «fewer|less [nn2*]» (fewer/less + plural noun):

    fewer benefits 50
    less benefits 13
    fewer problems 111
    less problems 15
    less taxes 49
    fewer taxes 14

    However, a search for «[mc*] [nn2*] or fewer|less» (cardinal number + plural noun + or fewer/less ) shows a fairly equal distribution.

  45. Peter Taylor said,

    December 3, 2010 @ 9:29 am

    @Pflaumbaum, as various people had already commented, the double "fewer" is inelegant. Replacing with less would make it worse: at that point I would consider it marked. I could speculate on whether I'm unusually inclined to "no more than" and why, but I'm sure Prof. Pullum would disapprove of that…

  46. Russell said,

    December 3, 2010 @ 10:02 am

    @PaulAtNorthGare

    Imagine: A strict follower of the discrete/continuum-based rule is working on a party platter containing piles of kernels of corn and shelled green peas. She counts all the kernels and peas and says,

    "Based on my calculations, the platter needs fewer corn(s)."

    Seems highly unlikely. "Fewer kernels," sure. "Fewer peas," yes. "Less corn," absolutely. And the prescriptivist rule would, if I understand it correctly, recommend "less corn."

    Sure, there's a general trend for discrete things to be denoted by count nouns and continuous things, by non-count nouns. But I'd say it's guided mostly by everyday human experience and understanding (now or in the past), not the latest quantificational technology. And there's always room of plain old lexical idiosyncrasy. And so the canonical fewer/less distinction is based on the lexicon's implementation of that conceptual difference, rather than an extra-linguistic determination of what counts (ahem) as what.

    (One could, if one wanted, create a distinction orthogonal to count/mass: pluralizable/nonpluralizable, and base less/fewer on one or both of those dimensions. Then "day" would be a mass, pluralizable noun, and "corn" would be a count, non-pluralizable noun.)

  47. Megs said,

    December 3, 2010 @ 10:13 am

    @Andrew

    "The asymmetry comes up with "number" and "amount" as well: "A number of people" is common, but "an amount of people" is pretty frequent as well, while "an amount of knowledge" is common, but "a number of knowledge" is nonsense."

    Ah, but people is a noun that is EITHER collective or plural, so it can take either the collective or the straight plural modifier. And that is what James pointed out too, and I think he's right. If we are using the implied subject of a collective OR something that can be in partial numbers, we are more likely to apply the collective adjective.

  48. Amir Khalid said,

    December 3, 2010 @ 10:18 am

    If the objection is to the ugly echo of "fewer" at the beginning and end of the sentence, rather than to any misuse of the word itself, then I suppose one could rephrase to

    Fewer than half of them have been decided by not more than ten points.

    How's that?.

  49. GeorgeW said,

    December 3, 2010 @ 10:22 am

    @Peter Taylor: I don't find the “double fewer” a problem (in its doubleness). I could revise the statement with two count nouns and the double 'fewer' sounds OK to me (ancient speaker of Southern AmE):

    'Quite often, these games don't even turn out to be good: Fewer than five of them have been decided by fewer than 10 points.'

    The "double fewer" seems even more acceptable with two unequivocal count nouns:

    1. Fewer than ten men came to the lecture while fewer than twenty women attended.

    2. *Fewer than ten men came to the lecture while less than twenty women attended.

  50. MJ said,

    December 3, 2010 @ 10:42 am

    @PaulAtNorthGare Time may be continuous, but the fact is we divide it up into discrete units. Indeed, a certain breed of metaphysical antirealist will tell you we divide up the items at the grocery store into discrete items as well.

  51. Megs said,

    December 3, 2010 @ 11:21 am

    @GeorgeW

    "The "double fewer" seems even more acceptable with two unequivocal count nouns:

    1. Fewer than ten men came to the lecture while fewer than twenty women attended.

    2. *Fewer than ten men came to the lecture while less than twenty women attended."

    This is also a case of parallelism. ALL things sound better if they can be made properly parallel.

  52. Pflaumbaum said,

    December 3, 2010 @ 11:24 am

    @ Chris, Tom Saylor –

    Certainly GKP's posts are frequently peevish, not to say bloody pissed off, and there may be little emotional difference between that feeling and prescriptivist peevishness.

    But the two aren't equivalent, because their cause isn't equivalent. One is rational and the other is irrational. GKP is a scientist who has to contend daily with ignorant people insisting on the teaching and application of laws that were largely formulated over a century ago and bear little relation to reality. His is, I'm guessing, the same rage as that of a climatologist confronted with climate change sceptics, except that 'language change sceptics' remain in positions of authority and virtually dominate the popular discourse.

    Imagine that many or most people still advocated and practiced the mediaeval theory of humours, without offering any evidence, or troubling to find out about modern medicine. A lot of doctors would be bloody annoyed about it. Their anger might mirror, emotionally, that of a quack faced with a patient who refused to accept bloodletting by leeches. But the differences is that the qualified doctor would be right.

  53. GeorgeW said,

    December 3, 2010 @ 11:31 am

    @Megs: Are you saying that the "double less" works only because of parallelism?

    How about: 'Fewer than a hundred fans came to the game in which fewer than five goals were scored.'

  54. Megs said,

    December 3, 2010 @ 11:35 am

    @GeorgeW

    Not at all.

    I'm saying the example you provided has less to do with sounding better due to both being whole counts than having to do with both are phrased the same way and so should be modfied the same way. That's parallelism.

    'Fewer than a hundred fans came to the game in which fewer than five goals were scored.'

    The above sentence is NOT parallel, but it's perfectly fine as written. Like I said earlier, I USE the rule of fewer/less and this doesn't phrase it in that ugly, weird sounding:

    A hundred fans or fewer came to the game in which five goals or fewer were scored.

    (Obviously that is a different meaning, but my point was that the phrase itself sounds wrong, whereas using "fewer than" sounds perfect.)

  55. Megs said,

    December 3, 2010 @ 11:38 am

    How about: 'Fewer than a hundred fans came to the game in which fewer than five goals were scored.'

    This particular example you provided, by the way, DOES illustrate your point.

  56. Megs said,

    December 3, 2010 @ 11:39 am

    The other didn't because of this:

    "The "double fewer" seems even more acceptable with two UNEQUIVOCAL count nouns:"

    emphasis mine.

    The sentence provided had EQUIVOCAL count nouns.

  57. GeorgeW said,

    December 3, 2010 @ 11:48 am

    @Megs: Thanks for the clarification.

    I suspect the ugliness of the "five goals or fewer" relates the 'X or less' construction that is so common without regard to count or mass.

    As I proposed earlier, I think 'less' is the default when there is any possible ambiguity.

  58. GeorgeW said,

    December 3, 2010 @ 11:58 am

    @Megs and Peter:

    Maybe I am missing a point. I might be arguing against something that wasn't claimed. I understood that proposed problem with a "double fewer" was due to the repetition. My examples were to challenge a repetition constraint.

  59. Paul Mulshine said,

    December 3, 2010 @ 12:16 pm

    What I've always thought particularly silly is when a prescriptivist uses the phrase "fewer than one." It violates the prescriptivist's own rule that only discrete quantities require "fewer" and quantities below one are not discrete since they by definition do not exist. "Less than one" not only sounds better but it makes more sense, since one is the first discrete quantity.

  60. Pflaumbaum said,

    December 3, 2010 @ 12:17 pm

    The repetition alone can't be the problem, as there's nothing wrong with

    Less than half of them have been decided by 10 points or less.

    It's the repetition in concert with the pedantry. The repetition calls more attention to the unnatural fewer.

  61. mollymooly said,

    December 3, 2010 @ 1:28 pm

    "Four and twenty virgins came down from Eskdalemuir…"

  62. Megs said,

    December 3, 2010 @ 2:08 pm

    Upon further reflection, I think the real problem with that sentence is that "Fewer than half" is ungrammatical! Yes, we can argue that games (which is what half modifies) is a plural noun, but the truth is "half" is not. It's a collective.

    The sentence would be perfectly grammatical by prescriptive grammar if it read:

    Quite often, these games don't even turn out to be good: Less than half of them have been decided by 10 points or fewer.

    This also illustrates where prescriptive grammar falls down. It reads much better and more clearly as:

    Quite often, these games don't even turn out to be good: Fewer than half of them have been decided by 10 points or less.

    This is because the point of the sentence is the games. Less calls less attention to itself and fewer takes focus. In the grammatically correct sentence, it makes the points seem like the focus and so the real topic is lost or at least muddied. In the second sentence, it makes a phrase of the stat instead of words and so the focus is that very few games have been good.

    Methinks I'm starting to confuse myself, but to finish the thought here:

    In cognitive linguistics, they ran a study and figured out that humans parse language in as few units as possible, meaning that we listen to phrases and expressions and hear the meaning of the entire phrase wherever possible. It takes more work for the brain to parse and combine individual words, so if opt to use stand-alone words instead of phrases, those words become subtly emphasised to the hearer/reader. In this case, it makes better sense to use the implied subjects, rather than the actual subjects, to select which modifiers to use.

  63. Allen said,

    December 3, 2010 @ 3:49 pm

    On the countable/uncountable distinction, are there "fewer," or are there "less" real numbers than there are elements of the set of all subsets of the real numbers?

  64. Peter Taylor said,

    December 3, 2010 @ 3:50 pm

    The repetition alone can't be the problem, as there's nothing wrong with
    Less than half of them have been decided by 10 points or less.

    IMO that's worse than the original sentence. Pace Megs,

    Less than half of them have been decided by 10 points or fewer.
    is an improvement.

  65. Pflaumbaum said,

    December 3, 2010 @ 4:02 pm

    There's no further we can go then. If two people with essentially the same dialect have completely opposite intuitions, then it's just a contested area.

  66. PaulAtNorthGare said,

    December 3, 2010 @ 5:51 pm

    @MJ

    Time may be continuous, but the fact is we divide it up into discrete units.

    No. What we do is label equally-spaced markings along a continuum. The surface text looks to be the same kind of thing, but the underlying concept is very different from, say, items at the checkout. You can have 2.3871 days. You can even have pi days. The difference between truly discrete quantities, and evenly-marked steps on a continuum – which merely use the same morphological markings as truly discrete quantities, but are conceptually very different – is a key one, and it's why I think this is much more a mathematical issue than a purely linguistic one.

    @Russell

    Imagine: A strict follower of the discrete/continuum-based rule is working on a party platter containing piles of kernels of corn and shelled green peas. She counts all the kernels and peas and says,

    "Based on my calculations, the platter needs fewer corn(s)."

    Seems highly unlikely. "Fewer kernels," sure. "Fewer peas," yes. "Less corn," absolutely. And the prescriptivist rule would, if I understand it correctly, recommend "less corn."

    I'm not sure if we're agreeing or disagreeing. I don't see a conflict between what you've described above and the count/mass heuristic. It's simply the case that the same stuff – and the same quantity of stuff – can, according to context, sometimes be thought of as both countable and mass. If I pour myself a glass of water, that's a mass, so I'd talk about "less water". But a scientist might, in context, talk about there being "fewer water molecules" in my glass than another. Water is an extreme example, because the component parts (the molecules) are so small, and there are so many in any real-world quantity, but the same applies to corn and to peas. We might, in context, talk about "fewer corn kernels", or "fewer peas", but at the same time we normally think of corn and peas as mass quantities, and the language we use changes accordingly.

  67. Megs said,

    December 3, 2010 @ 6:22 pm

    I think of peas as countable, but corn? Please!

    Less corn. Thank you.

  68. MJ said,

    December 3, 2010 @ 8:15 pm

    @PaulAtNorthGare Surely you're not suggesting that English speakers have been wrong in speaking of fewer hours, fewer minutes, fewer days, fewer weeks, etc., for all these centuries?

  69. PaulB said,

    December 3, 2010 @ 9:09 pm

    Checking what sounds right and wrong to me, I find that I prefer "fewer" for anything that's counted, but I interpret that narrowly: it applies only to natural numbers (1,2,3…) of discrete things. "Seven days or less" is correct for me unless I'm explicitly counting the days.

    That makes the first "fewer" in the WSJ sentence wrong. "Fewer than half" is disallowed. I suppose one might parse the sentence as meaning "Fewer than the natural number found by taking half the games and rounding up if necessary" but no one's real-time grammar checker is going to do that (I hope).

    The second "fewer" is marginal for two reasons. First, because points in American Football are scored mostly three or six at a time, and are therefore added rather than counted. And second, because point differences can be negative and so are not generically natural numbers. (One would not describe a team in an Association Football league as having "a goal difference of fewer than ten".) In the WSJ sentence it is the size of the difference that's being discussed, which is always positive, but again that's rather too much to process by ear.

    The big problem with the sentence is that the repetition of "fewer" draws our attention to the rather unnatural choices being made. It's not much improved by changing the first instance to "less" because one's attention is then drawn to the fine distinction being made. "Less" in both places is preferable. But it would much better to avoid the double inequality by writing "Most of them have been decided by 10 points or fewer" or "Most of them have been decided by 10 points or less".

  70. GeorgeW said,

    December 3, 2010 @ 9:29 pm

    Based on several comments here and my own intuition, there seems to be a difference between "less than' and "or less" independent of the noun. Both of the following work for me:

    1. Fewer than 10 books.
    2. 10 books or less.

    But not:
    3. *Less than 10 books

    Seems OK, but a little stiff:
    4. 10 books or fewer.

  71. PaulAtNorthGare said,

    December 3, 2010 @ 10:23 pm

    @MJ

    Surely you're not suggesting that English speakers have been wrong in speaking of fewer hours, fewer minutes, fewer days, fewer weeks, etc., for all these centuries?

    Sometimes, but not always. And I realize that what I wrote wasn't clear. Aside: I don't really mean to suggest that anyone is wrong, except maybe in how they see mathematical concepts sometimes. My main point is that for me the count/mass heuristic works extremely reliably as a guide to whether I'll prefer "fewer" or "less", and I think it's closer to how people choose between the words than they sometimes are aware of. The key is understanding exactly which measures are count/discrete, and which are mass/continuous.

    I can imagine situations in which "hours" (and "minutes", "days", "years", etc.) are indeed count/discrete. I'd argue that they're less common, though. For example, propose a timetable of classes which is divided formally into hour-long slots. A teacher might then say something like: "I can only teach three hours or fewer". Here, "hours" is somewhat synonymous with "classes", and they genuinely can't be divided. The point is that they're counting slots, and not measuring time.

    This is a quite different situation from one in which the same teacher might say: "I can only stay behind after classes for three hours or less" (clumsy, but you get the point). Here, continuous time is being measured, rather than slots being counted, so my heuristic prefers "less".

    So, when someone uses a phrase like "three hours", it's completely true that, in context, they might be either counting discrete slots ("three hours or fewer") or measuring continuous time ("three hours or less"). They're not interchangeable, though, and the latter is much more common.

    If there's something here I'm calling "wrong", it would be using "three hours or fewer" on the assumption that discrete hours are being counted, when in fact what's happening in that situation is that continuous time is being measured. But it's more a mathematical mistake than a linguistic one.

  72. Ellen K. said,

    December 3, 2010 @ 11:16 pm

    @GeorgeW

    "Less than 10 books" works fine for me. I actually prefer it to "Fewer than 10 books", which comes across to me as more formal, and if I imagine it spoken, sounds rather stuffy.

  73. ignoramus said,

    December 3, 2010 @ 11:20 pm

    "…Fewer than half of them have been decided by 10 points or fewer…"
    surely it could be written that:
    Less than 50 percent of the games had scores under 11points.

    by one that failed 11+

  74. Ellen K. said,

    December 3, 2010 @ 11:22 pm

    One more note. If I were an ESL teaching teaching these words, I'm thinking one thing I'd include is "When in doubt, use less".

    Oh, and another note on books… without a number, I think "fewer" does sound better, though "less" isn't wrong. No * for "less books".

  75. Atmir Ilias said,

    December 4, 2010 @ 1:36 am

    I think the "less" contains the tendency of movement toward negative direction and the "fewer" contains a quantity of numbers moving toward negative direction. "Less” is like to say "moving from a given number toward negative direction, no matter how much". "Fewer” is like to say " moving from a given number discounting a small quantity of numbers toward negative direction.

  76. Bob Ladd said,

    December 4, 2010 @ 2:58 am

    As several people have suggested, part of the problem with the prescriptivist rule – and this is what GKP was getting at with his "7 days or fewer" example – is that the clear grammatical distinction between count and mass has to be applied to a world in which there are borderline cases. Quantities of time, money, and points in a game are such borderline cases. So are things that come in quantities of hard-to-count individual units, like peas and corn.
    One way to see this clearly is to compare across languages. In our bilingual Italian/English household we bump into this regularly with hair (count/plural in Italian, mass/singular in English), grapes (mass/singular in Italian, count/plural in English), and various others. Same with money – in Italian you normally say the equivalent of Ten euros are too much to pay for that or There are 50 euros in my wallet, whereas in English those are far more natural with singular verbs.

    The grammar of both languages (and the prescriptivist rule about less and fewer) gives you an either/or choice between singular and plural, mass and count. In most cases, it's no problem to assign whatever you're talking about to one or the other. But not all.

    For what it's worth, my usage of less and fewer generally follows the prescriptivists' rule, but I agree with GKP that the WSJ sentence is horrible. Peter Taylor's edit is great.

  77. tpr said,

    December 4, 2010 @ 5:10 am

    Part of the trouble stems from relying too heavily on semantic cues. A good example is the word spaghetti, which is a singular mass noun in English ("Some spaghetti is…"), but was borrowed from Italian where it's a plural count noun ("Gli spaghetti sono…"). It refers to the same kind(s) of entity in the two languages, so semantic cues aren't sufficient to determine which grammatical category the words belong to in either language. As with grammatical gender, semantic cues help you to guess, but they're not enough. In particular, there is a grey area with countable objects that appear in a form that we're more inclined to measure than count even if there are countable objects to be found at a fine granularity (e.g., rice, data).

    Another part of the trouble is that less doesn't map exclusively onto mass nouns, even in the register of English that permits fewer (which does appear to map exclusively onto plural count nouns). Less is acceptable with a range of plural count nouns that refer to quantities on a scale (e.g., less than 3.25 points / degrees / euros, but not less than 3 apples at least in the register of English that uses fewer).

  78. GeorgeW said,

    December 4, 2010 @ 6:18 am

    @Ellen K: I think you are right – "less books" would not be universally ungrammatical.

    One more point to illustrate the general 'less' preference over 'fewer.' We have the fixed expression, 'more or less' irrespective of the referenced noun. Example:

    *There are 10 discrete, countable units on the table, more or fewer. (Ugh!)

  79. Julie said,

    December 4, 2010 @ 6:45 am

    In my normal conversational register, I have no particular use for "fewer." If I were to use it, I would be being hyper-precise and much more formal than I usually am, even in formal contexts. In writing, I apply the prescriptivist rules, more or less….I trust my "mind's ear" far more than any hard and fast rule. That means I use "less" somewhat more often than the prescriptivists would have me do.

    "Quite often, these games don't even turn out to be good: Fewer than half of them have been decided by 10 points or fewer. " I don't care about the first "fewer," that's just the WSJ being as formal as they possibly can. The second one strikes me as wrong.

    At any rate, the whole sentence is awful, right down to its underlying concepts. A game decided by less than 10 points is thereby "good?" Perhaps it should be more like "Fewer/less than half of them were close enough to be exciting."

  80. tpr said,

    December 4, 2010 @ 6:48 am

    @GeorgeW
    "More or less" is idiomatic in that context though. In a context where it isn't idiomatic, it sounds fine:

    Are there more or fewer than 10 books on the shelf?

  81. John Ferguson said,

    December 4, 2010 @ 7:17 am

    Had the estimable Mr Pullum merely stated that the construction used by the WSJ sounded bad when read aloud and that the subeditors just needed to take a chill pill and let things slide, I would have been in agreement. Why should the quality of prose suffer because of one teensy grammar rule, especially in an headline? But… Mr Pullum went on such a tirade that I instinctively found myself disagreeing with him on the broader argument. Why not be more prescriptive? English is a rather malleable language and promoting descriptivism over prescriptivism means that the language becomes ever harder to learn and continues to be a moving target so that instead of a few rules and some experience, a new learner needs extensive experience to grasp the ever greater number of special cases.

    Further, the word less actually means small, e.g. the disciple known as James the Less for the KJV readers, and hence should not be used as a comparative at all! (Insert green text below on how less became a comparative)

    Not being an ornithologist, it was only a few years ago that I even realised the phrase 'lesser spotted', as applied to birds, actually means 'smaller and with spots', not that it is seen on fewer occasions than another type. So I contend more prescriptivist education is required.

  82. GeorgeW said,

    December 4, 2010 @ 8:22 am

    @John Ferguson: I do not agree that presciptivism promotes better learning for either first or second language learners.

    Children have no difficulty acquiring the language to which they are exposed. It is when they go to school that they are taught another form.

    I can see no reason why 2nd-language learners would benefit from presciptivism rather than being taught the language that is actually spoken.

    As to the less/fewer distinction, I think a good case can be made that language acquisition or learning would be enhanced were there not two forms to deal with. My experience with ESL speakers is that this is a difficult rule to acquire. I think this is quite marked cross-linguistically. I am not sure if any other (at least non-Germanic) language makes this kind of distinction.

  83. Pflaumbaum said,

    December 4, 2010 @ 8:48 am

    @ John Ferguson

    That would be fine if the prescriptive rule you were talking about were actually valid. There are good grounds for ensuring that children learn standard rather than non-standard dialects, for situations like exams and job interviews that require it.

    But using less with count nouns isn't non-standard (let alone ungrammatical). It's just used less often in formal contexts than it is more generally. And there are lots of educated standard English speakers (like me) who barely use fewer at all.

  84. John Cowan said,

    December 4, 2010 @ 11:58 am

    John Ferguson: I can't engreen my text, but here's the story with less: it began as a comparative, specifically the comparative of little. So St. James the Less was so called because he was less (smaller, younger, less important — nobody knows which) than St. James the Great(er). Nowadays we have regularized this comparative and would call him "St. James the Lesser" instead, as in the Lesser Spotted Woodpecker. Similarly, least is the old superlative, and that shows up in the name of the Least Tern.

  85. F. Escobar said,

    December 4, 2010 @ 5:32 pm

    "a nation of cowed and abused users of the English language, constantly fearing punishment for violation of crassly inaccurate and simplistic rules that they do not understand, rules enforced by stupid style manuals, ignorant writing instructors, dimwitted editors, and their cowardly henchmen." This is delightfully harsh. But I wonder, dejectedly, what GKP would think of speakers of Spanish, whole masses of whom constantly fear a thunderbolt hurled by the Royal Academy for misplacing an accent mark.

  86. John Ferguson said,

    December 4, 2010 @ 8:57 pm

    Kindest regards due to Mr Cowan.

    Pflaumbaum: Indeed, children pick things up very well, but they also form egg-corns and faux-grammars very easily. Allowing the odd mis-spelling or grammatical inconsistency just lets everyone get on with their lives, but it would be a disservice to all learners not to strive for a little standardisation.

  87. Pflaumbaum said,

    December 5, 2010 @ 6:18 am

    But the fact remains that this isn't an eggcorn or error or deviation from standard dialect. It makes sense to remind children that, say, "I ain't seen nothing", while not ungrammatical, will fail to impress at a job interview. But what are the grounds for 'standardisation' of the less/fewer restriction? If there is to be standardisation, shouldn't it be in the direction of the majority who don't tend to replace less with fewer?

  88. Tom Saylor said,

    December 5, 2010 @ 7:11 am

    To Pflaumbaum, who said:

    "Certainly GKP's posts are frequently peevish, not to say bloody pissed off, and there may be little emotional difference between that feeling and prescriptivist peevishness. But the two aren't equivalent, because their cause isn't equivalent. One is rational and the other is irrational. GKP is a scientist who has to contend daily with ignorant people insisting on the teaching and application of laws that were largely formulated over a century ago and bear little relation to reality. "

    So what you're saying is that no prescription is prescriptivist per se? Whether a prescription is prescriptivist or not depends on who's doing the prescribing and whether that person's "cause" is rational or not? It's possible, then, in principle, that a given piece of usage advice is both prescriptivist and not prescriptivist if it happens to be endorsed by both a good guy and a bad guy, someone with a rational cause and someone with an irrational cause?

  89. Pflaumbaum said,

    December 5, 2010 @ 8:49 am

    I'm saying the emotion expressed might be similar. They both qualify as peevish, but they don't both qualify as prescriptivist. Just as in my medical analogy, both doctor and quack might have tantrums, but only the quack's tantrum is in defence of superstition.

    The difference between medicine and superstition is established by experiment and peer review, so yes it depends whose cause is rational and whose ir-, but that decision about rationality is not arbitrary, as you seem to be implying. The quack, like the prescriptivist, may well believe his theory to be rational, but that oughtn't to count for much unless it can be properly justified.

    It's true that when GKP has a go at some inelegance resulting from slavish adherence to a non-existent grammatical rule, he is to some extent telling the author how to write. But descriptivists have never waived the right to an opinion on what's good use of language and what isn't. What separates them from prescriptivists is their insistence that the facts of a language – like its grammatical structure – can only be gleaned from observation and theory, not declared by fiat.

  90. GeorgeW said,

    December 5, 2010 @ 9:26 am

    @Tom Saylor: It is one thing to 'describe' how a language IS spoken and quite another to 'prescribe' how it SHOULD be spoken.

    Further, we should not confuse register with 'proper' language. The fact that a particular form is more appropriate in a formal context does not entail that it should be the form in all contexts.

  91. Army1987 said,

    December 5, 2010 @ 12:03 pm

    The asymmetry is in part due to the fact that natural numbers are [isomorphic to] a subset of the real numbers, but not vice versa. (But I don't think it's *only* semantic: "too much beans" sounds weird even if what you're talking about is their total mass rather than the number of individual beans.)

  92. Bev Rowe said,

    December 5, 2010 @ 6:44 pm

    I'm sure people have pointed this out but never understand the "logical" explanation for the difference between "lesser" and "fewer" when "more" works with no such distinction

  93. Dan K said,

    December 5, 2010 @ 9:22 pm

    Here's the simple test for whether a strongly expressed opinion on what's right or wrong is prescriptivist poppycock or sage advice. It's all in the why. For true prescriptivist poppycock, the why is almost invariably "because that's the rule." For sage advice, the whys are more varied, but usually some variant of, "because there's an actual reason why it's bad (or good), and this is what that reason is." Of course, there are ways to dress up poppycock, so sometimes you have to disassemble it a bit, but that's what Language Log is for, right?

  94. Tom Saylor said,

    December 6, 2010 @ 12:23 am

    To GeorgeW, who said

    "@Tom Saylor: It is one thing to 'describe' how a language IS spoken and quite another to 'prescribe' how it SHOULD be spoken."

    Certainly, and GKP does both. My point is that his prescriptions (explicit and implicit) are not different in kind from the prescriptions of those he derides as prescriptivists. They are, at bottom, denunciations of usage that he personally finds inelegant or otherwise offensive, just as the prescriptions of Fowler and Simon are basically just denunciations of usage that they don't like.

  95. tpr said,

    December 6, 2010 @ 6:09 am

    @Tom Saylor
    The difference between descriptivism and prescriptivism isn't entirely about is versus ought. The more important difference is about where each believes the 'rules' of language come from.

  96. Tom Saylor said,

    December 6, 2010 @ 8:32 am

    To tpr:

    The "rule" against saying "10 items or less" comes from some long dead grammarian who preferred "10 items or fewer." The "rule" against saying "10 points or fewer" comes from the likes of GKP, who prefers "10 points or less." Both of the proscribed usages are abundantly attested in the writings of English speakers. So why is the one proscription deemed prescriptivist and the other not?

  97. tpr said,

    December 6, 2010 @ 9:47 am

    @Tom Saylor
    I can't speak for GKP, but I think you're failing to distinguish prescriptive statements about usage with 'prescriptive' statements about prescriptivism itself. GKP appears to believe that evidence should be the standard when people make statements about what the rules of grammar are. That would make him descriptivist about usage, but not about the behaviour of people who say unsubstantiated things about language. You could use the word 'prescriptivist' in a broad sense to encompass this stance, but that would obscure an important distinction that is really there and present a false equivalence between the two sides.

  98. Sam said,

    December 6, 2010 @ 10:03 am

    @Tom Saylor,
    You seem very keen to label GKP a prescriptivist and determined to miss the point of the sound arguments presented above (and repeated many times elsewhere in LL) in your pursuit of this goal. GKP is not prescribing a rule. Rather he expresses exasperation at those who prescribe a 'rule' by diktat – and the clumsy sentences produced by those attempting to follow this non-rule – that does not match the usage of most people (and never has). In so doing he does not definitively prescribe an overarching rule as those deemed 'prescriptivist' do; rather, he notes standard use (that is, what most people usually say and write). On this score you might want to read the later post by Mark Liberman.

  99. Ellen K. said,

    December 6, 2010 @ 10:42 am

    @Tom Saylor: As I understand it, Professor Pullum isn't proscribing "10 points or fewer". Rather, based on experience with the Wall Street Journal, and based on usage patterns ("10 points or fewer" being the more common way to write it), he concludes that the Wall Street Journal writer and/or editor chose "or fewer" because of a misguided idea that "or fewer" is preferable here. And observing that (in his opinion) this makes the passage sound bad.

    I suppose you could call him a proscriptivist of a sort, rather than a prescriptivist — he's proscribing makes language choices based on rules with no basis in actual usage.

  100. Tom Saylor said,

    December 6, 2010 @ 12:25 pm

    To Sam and Ellen K, who contend that GKP is not proscribing usage.

    Prescriptive grammarian do not always explicitly proscribe the usage they wish to condemn. Their proscriptions are, as often as not, implicit in their characterizations of the usage. Rather than saying "Don't do X" or "One shouldn't do X" they often say "X is wrong" or "X is unacceptable" or "X is abominable," clearly implying a proscription on the use of X. So when GKP characterizes the use of "10 points or fewer" as *grossly unacceptable*, I take him to be proscribing that usage. Is that not a fair inference to draw? Moreover, GKP simply assumes that in saying "10 points or fewer" the author of the article (or his editor) is bending to a prescriptivist rule rather than following his own linguistic inclinations. And yet the linguistic evidence quite clearly shows that many, many people are inclined to write things like "10 points or fewer" of their own accord–that this usage is widely regarded as acceptable. So what do you call it when someone implicitly proscribes (as "grossly unacceptable") a usage that many speakers of the language feel is perfectly acceptable?

  101. Ellen said,

    December 6, 2010 @ 12:44 pm

    Actually, it's the full sentence with the double "fewer" that's "grossly unacceptable".

    But, perhaps more importantly, I take such colorful language to be an expression of personal opinion, rather than an equivalent of "don't do that".

    And do you really have evidence that people use "10 points or fewer" without it being a case of trying to follow a prescriptivist rule? I don't even know how you could know if that's the case or not.

  102. CliveStaples said,

    December 6, 2010 @ 1:10 pm

    Erm, isn't it prescriptivist to say that you shouldn't say "10 points or fewer"? Shouldn't usage dictate?

  103. tpr said,

    December 6, 2010 @ 1:19 pm

    when GKP characterizes the use of "10 points or fewer" as *grossly unacceptable*, I take him to be proscribing that usage. Is that not a fair inference to draw?

    Not a fair inference no. 'Acceptability' in linguistics has a technical meaning that differs from what you've taken it to mean here. He's not making a judgement about what people should say. Descriptive linguists routinely use what they call 'acceptability' judgements of native speakers to investigate the nature of the mental representations at work. An 'unacceptable' utterance in this sense is just one that native speakers consider grammatically anomalous. In this case, by "grossly unacceptable" he was most definitely talking about utterances that native speakers consider grossly anomalous. I can see why that would be confusing though.

  104. Tom Saylor said,

    December 6, 2010 @ 3:32 pm

    To tpr:

    If you’re right–if GKP is using “unacceptable” in some specialized sense to indicate that the “10 points or fewer” construction is deemed ungrammatical by the vast majority of educated speakers of English–then he is exonerated of the charge of prescriptivism but guilty of bad descriptivism. A few simple Google searches will plainly show that this construction is very widely accepted in the English-speaking world.

  105. Andrew (not the same one) said,

    December 6, 2010 @ 4:45 pm

    I think that descriptivists do sometimes appear to be describing, not the language as it actually is, but the language as it would be if those nasty prescriptivists did not get in the way. I see statements like 'this rule does not accurately describe the usage of speakers, except those who are speaking that way because of the rule', the assumption being that this last group does not count.

    This does not seem right to me. Language is what it is because of the social forces acting on it, one of those forces being prescriptivism. Some usages that originated because of prescriptivism are quite widespread, and an accurate description of usage should therefore recognise them. So even if 'ten points or less' is what rises naturally to the lips of the untutored speaker, enough people have had 'fewer with a count noun' dinned into them in childhood that 'ten points or fewer' is now also a possible usage.

  106. Pflaumbaum said,

    December 7, 2010 @ 5:18 am

    I think GKP and other descriptivists do recognise that. A case in point is nominative co-ordinated pronouns in object position, as in:

    %The doctor gave it to Liz and I

    The CGEL describes this as grammatical in some standard dialects, even though it has only arisen through (confusion about) prescriptive rules against accusative pronouns in co-ordination. It is so widespread that it has become standard (though contested).

    The case with Ten points or fewer is quite different, as the data in Prof. Liberman's later post showed:

    http://languagelog.ldc.upenn.edu/nll/?p=2826

    If it became as common as Tom is claiming, I don't doubt that would be recognised. But it's not the case that if a small minority of people has a particular usage, descriptive grammarians are obliged to describe it as grammatical in standard English. Even if that group contains lots of sub-editors.

  107. Tom Saylor said,

    December 7, 2010 @ 6:16 am

    To Pflaufbaum, who seems to contest the claim that the "10 points or fewer" construction is widely accepted among English speakers:

    The claim you contest is actually *supported* by the LL post to which you point. Mark Liberman there reports:

    "As for things like votes and points, we can find plausible examples that go both ways, in approximately similar numbers, from well-written and well-edited sources like the New York Times"

    By "go both ways, in approximately similar numbers" he means that in the examples studied the "10 points or fewer" construction is about as frequent as the "10 points or less" construction. Is that not evidence that the "10 points or fewer" is widely accepted?

    :

  108. Pflaumbaum said,

    December 7, 2010 @ 7:13 am

    You're right. I should have said numerically-quantified plural noun phrases in general, it seems to be more acceptable with points and votes. I still don't agree that GKP was being 'prescriptivist', but he may have underestimated the extent to which this particular prescriptive rule has changed behaviour.

  109. Tom Saylor said,

    December 7, 2010 @ 10:33 am

    To Pflaumbaum:

    You and I both seem to be applying the principle of charity to GKP’s comment, but in different ways. You find it unthinkable that GKP could be prescriptivist, so you charitably interpret his “grossly unacceptable” comment as a bit of mistaken descriptivism. I, on the other hand, find it difficult to believe that GKP could have made such a mistake in describing actual usage, so I charitably interpret the “grossly unacceptable” comment as prescriptivist. That’s not intended as an insult. Unlike many people here, I don’t use “prescriptivist” as a term of reproach.

  110. Robert said,

    January 4, 2011 @ 4:16 am

    Is there any one else out there who finds the "fewer than half" more annoying than the "fewer than 10 points"? I think my feeling about this is related to the point raised earlier about the distinction between discrete and continuous variables.

    I can't instinctively accept that "a half" is countable. It's a measure or statistic, and even when a number of games are concerned it may have a fractional part.

  111. Jason said,

    April 12, 2011 @ 10:24 am

    I think 'fewer' could in a strict sense be thought to violate the Rule (the sacrosanctness thereof demanding the dignity of the uppercase) in this case. Essentially (in the Aristotelian sense) we aren't talking about points we are talking about scores "by (a score of) 10 points or…" scores are what decide games not points.

  112. Edward said,

    August 23, 2011 @ 3:16 am

    I don't find it ridiculous in the lease. I schooled under the British system and to me "10 items or fewer" sounds correct and natural, and "10 items or less" sounds quite odd.

    However, when it comes to things like time, "10 hours or less" sounds correct because of ellipsis. It is an abbreviation of "10 hours' time or less".

    Here in the States, the distinction between "formal" and "casual" English is a bit difficult for me to get used to. During my education, there was only proper and improper, and the latter was discouraged, not embraced as a viable alternative.

    So, of course, I would never say "fewer beer", but nor would I EVER say "less beers" … unless perhaps I'd had quite a few to many.

  113. Edward said,

    August 23, 2011 @ 3:17 am

    doh … *too

  114. Edward said,

    August 23, 2011 @ 3:23 am

    Also, I submit that "fewer than half" is actually grammatically incorrect.

    "Less than half", is an ellipsis for "less than half the number of".

  115. Gavin Hodgkinson said,

    June 19, 2014 @ 11:06 am

    To put less than two shots of whisky in a glass would be to diddle the customer. To put fewer than two shots of whisky in a glass would be to put one in it. The customer would notice.

    ‘Shot’ is a divisible unit. ‘Year’ is a divisible unit. ‘Item’ is not a divisible unit. That is why we can say 'Two shots of whisky or less', '5 years of age or less' but cannot say '10 items or less'. Pullum hasn't realised this.

    First he slurs King Alfred's good name and then goes on to admit that King Alfred did not use 'less' with a plural noun. The phrase in question is 'swā mid læs worda swā mid mā'. 'læs' is a noun here – an indefinite pronoun. Where there's a preposition, there's got to be a noun, right, folks? And 'worda' is genitive. Furthermore, 'læs' and 'mā' don't even seem to possess a comparative aspect. That would be 'læssa' and 'māra'. Don't nail me on this point yet; I will have to do more research. In essence the phrase means "either with a modicum of words or with plenty". Can anyone provide the original Latin phrase?

  116. Brian Hitchcock said,

    July 21, 2014 @ 5:59 pm

    All sub-editors are by nature over-picky. That is, they pick over things, like a scavenger. I believe the writer was concerned that the sub-editor apparently had been OVERLY PICKY.

RSS feed for comments on this post

Leave a Comment