Subject: “Ten items or less” is just fine
Okay, I’ve just violated a principle of suspenseful writing by distilling my entire message into my headline. But I’ll bet it made you pay attention, because if you’re like most editors, you have a definite opinion about this one.
In my October 2 blog post about William Safire, I mentioned that a fellow “Lexicographic Irregular” was going to publish a remembrance of Safire, and I wasn’t going to steal his thunder about a specific point he made in the article. But now that the article has appeared (”The Maven, Nevermore,” by Ben Zimmer, in The New York Times Magazine, October 5, 2009), I can rant: I dearly wish that William Safire had never decided that “Ten items or fewer” was the correct way to style the sign above a supermarket’s express checkout — at least not to the point that he then felt compelled to persuade the Safeway grocery chain to change its signs from the disdained “Ten items or less.” I’ve noticed that the Whole Foods stores have painstakingly followed suit, and that is the problem: Safire’s campaign granted undue legitimacy to a misconception, and now it’s almost impossible to persuade the sign makers and customer service people otherwise — even when the result is something as patently idiotic as the following message I get when I send an e-mail to Fresh Direct, a grocery-delivery company: “Thank you for your message. We’ll respond in 24 hours — or fewer.”
How “normal” do the following sentences sound to you?
If the temperature is 20 degrees or fewer, we will keep the children indoors for recess.
When the truck is carrying a full load, it gets 15 miles to the gallon or fewer.
Most of these devices now cost $50 or fewer.
Please summarize the key selling points in 100 words or fewer.
If you work three days a week or fewer, you are considered half-time.
I could go on and on. The reason your instinct is to change fewer to less in the above sentences (unless you’ve had your instincts drilled out of you by too many visits to an acquiescent supermarket) is that each of the amounts in the above sentences represents a benchmark. When an amount is perceived as a metaphorical line — something one can go over or come under — it is notionally singular, not plural, and so less should be used, not fewer.
We generally don’t make mistakes with rates (miles per hour, dollars per hour, days per week, and so on), and that’s because the expression “X per Y” is perceived as a total amount. The singular rate (per hour, per day, per week, etc.) also separates the word with the plural marker on it (miles, dollars, days, etc.) from the alternative of or fewer/less, which helps prevent false attraction to the plural. Some people begin to be hesitant when it comes to time, temperature, and money, but most of us still get it right. When it comes to things that fall outside the familiar categories (rates, time, temperature, money), though, we often stumble.
And that’s where the problem with items on a list comes in: if the word items is plural, why does “ten or fewer” sound wrong? It’s because “ten items” is also a benchmark. If you come under it, you get the speedy line. If you go over it, you’re stuck behind the family with two carts, even if you have only 11 items. The sign doesn’t necessarily mean “ten items, or fewer items than that.” It can just as easily mean “ten items, or less than this benchmark.”
I am not about to mount a counter-campaign to persuade Whole Foods to see the light. But I did write to Fresh Direct, on the grounds that the company might want to know that it had taken things a bit too far and was making its customers snort (because I just know I’m Everycustomer). So far, I haven’t heard back, and the response message is unchanged. To its customer service people, I’m clearly another crank; that, to me, is a permanent — if relatively unimportant — legacy from William Safire.
If you enjoyed this post, make sure you subscribe to our RSS feed!
Tags: English grammar, Less vs. fewer, Safire
[...] This post was mentioned on Twitter by Chris Davey, Copy Editing. Copy Editing said: New blog post: Copyediting Tip of the Week: "Ten items or less" is just fine http://bit.ly/1nt1bt [...]
Thanks for the voice of reason. Better tell Weird Al: http://www.twitvid.com/EFE89
While I agree with the fundamental argument you are making, I think fewer works in some of the sentences you’ve used as counter-examples to normalcy. However, the sentences should be slightly re-phrased:
When the truck is carrying a full load, it gets 15 or fewer miles to the gallon.
Most of these devices now cost 50 or fewer dollars.
Please summarize the key selling points in 100 or fewer words.
If you work three or fewer days a week, you are considered half-time.
And thus, the express line signage could work similarly:
10 or fewer items.
However, all that said, I’m merely a hack who feels compelled to share my 2 cents.
Strongly disagree with almost all of this. Notwithstanding that the less/fewer issue in supermarkets is a storm in a teacup, the most useful semantic difference between ‘less’ and ‘fewer’ has nothing to do with ‘benchmarks’ in the great majority of cases, but simply with the difference between discrete and continuous quantities. “…[T]emperature is 20 degrees or fewer” isn’t awkward because it’s a benchmark, but because temperature is continuous, so we’re not counting in whole degrees, and the same is true for money, distances, etc.
Items at the checkout, however, can’t be subdivided, so - descriptive/prescriptive arguments aside - ‘fewer’ *is* more appropriate given the typical meanings ascribed to ‘less’ and ‘fewer’. The discrete/continuous semantic distinction between them is a very useful one, and it would be a shame if it was lost or blurred.
The difference between less and fewer is actually one of discreteness. That is, one says twenty degrees or less because it could be, conceivably, 19.5, or 19.954, or 16 + pi degrees. Similarly, a trucks travels 15 or less miles per hour, or $50 dollars of less, because both mile per hour and dollar can be divided further and further. One has, for example, fewer pairs of socks, because one can divide a pair into at most two things before the concept becomes meaningless, whereas one can work three days per week, or just a tad less or a tad more. In fact, of the examples you provide at the beginning of your article, the discrete vs. continuous distinctions explains all but the “100 words or fewer” case, which is generally considered the correct form anyway.
[...] This post was mentioned on Twitter by Bill Walsh and Josh Neas, Edmund O’Neill. Edmund O’Neill said: I love how prescriptive language norms are slowly be eroded away by prescriptivists. This is the way it goes, folks. http://bit.ly/4shJAT [...]
Sorry, but count me in with Paula and Pavel. Benchmarking has nothing to do with the distinction.
I think it’s interesting that Pavel is using the idea of continuous versus discrete amounts to (mostly) agree with me, and Paula is using it to disagree. As you might expect, I side with Pavel.
To answer Paula: Degrees are discrete amounts by which we measure temperature. It’s the plural word “degrees” that tends to trigger the sense that “fewer” must be the right choice, but it’s not: “It’s less than 20 degrees outside” is a statement that the temperature is lower than a specific level, or benchmark—and Pavel is, I think, right to point out that it could be 19.5 degrees, and we’d never say “It’s 19.5 degrees or fewer” outside because we don’t know what increment we’re counting in anymore. The amount “10 items” can certainly be subdivided into each individual item, but I still say that the supermarket is setting a benchmark even though you can’t have, say, seven and a half items in your basket.
Pavel’s “pairs of socks” example is a red herring, because when you’re counting how many pairs of socks are in your drawer, you’re not setting a benchmark. We’re talking here about a specific context, that of notionally singular levels that one can go over or come under.
In that sense, SeaCloud’s suggestion brings a point that I think is crucial here. Moving “fewer” or “less” to a position in front of the noun—“15 or fewer miles to the gallon”—changes the grammar. “Fewer” ceases to be a pronoun and becomes a modifier (specifically, a type of determiner called a “quantifier”) that has to agree with the word it precedes. The pronoun can be singular because its antecedent is a notionally singular level, but the modifier doesn’t have that luxury: it has to be plural because the word it modifies is plural.
But applying this in statements of measurement or rate is generally considered to be stylistically awkward—when Pavel wrote “a truck travels 15 or less miles per hour,” it sounded wrong for a reason. The stated rate is “15 miles per hour”; a truck can get more or less (not fewer) than that rate. By introducing the modifier into the rate, the notion of the singularity of the rate is violated. Essentially you now have two competing grammatical notions: the deeply embedded sense that rates are singular concepts, and the deeply embedded pattern of making a quantifier agree in number with the individual word it modifies. Hence the awkwardness, and our preference for using the construction with the pronoun—“15 miles to the gallon or less.”
Where the idea of continuous versus discrete amounts falls down is precisely in the place where our disagreement still stands. Pavel says that “100 words or fewer” is “generally considered the correct form anyway,” which sidesteps the problem: the discrete vs. continuous explanation fails to account for “100 words or less” not because a lot of people who’ve been taught by Miss Thistlebottom think it’s wrong so it must be; it fails to account for “100 words or less” because that requirement is a benchmark, just like “10 items or less.”
I prefer “fewer” for discrete objects, but I can’t disagree with Wendalyn’s logical argument, which is supported by a tradition of broadly accepted usage. I would say it largely is a matter of perspective, and as the numbers get bigger, we are less likely to view objects as distinct. One hundred words or less seems more palatable than 10 items or less. No one counts characters when composing an item for Twitter; we’re told when we have passed the benchmark. Thus, one rarely says “140 characters or fewer.” This may make little difference to Wendalyn’s argument (or the opposite view). If you accept the number as a benchmark, less works for 10,000 items or 10 items. But if it’s something I might count, such as the items in my grocery cart, I’ll stick with “fewer.”
[A quick point of order: it's Paul, not Paula. Sorry for not making that clear. My screen name is Paul-at-North-Gare.]
Unless I’m misreading, I think Pavel and I are saying basically the same thing, which is that because the criterion for choosing between less/fewer is typically the distinction between discrete and continuous quantities, and not some abstract idea of a ‘benchmark’ having been set, you’re right some of the time, but for the wrong reason.
No, they’re not. They’re (largely arbitrary) points on a continuum, and the difference is a crucial one.
How would you support this assertion? I doubt that many people would find themselves tending towards ‘fewer’ with any continuous measure. The example you give (’24 hours - or fewer’) doesn’t strike me as a natural tendency, so much as confusion from having misunderstood a debate that wasn’t followed with any insight.
It’s not clear how you’re defining ‘benchmark’ here, and that might be part of the problem. Although disagreeing with your analysis, I can see how ‘10 items or less/fewer’ represents some sort of pass/fail threshold, which you might refer to as a ‘benchmark’. But ‘it’s less than 20 degrees outside’ is a simple statement about the temperature, without any sense of requirement, or pass/fail test. If ‘it’s less than 20 degrees outside’ fits within what you think of as a ‘benchmark’, then almost any comparison would fit, and the term would cease to have any meaning.
No. We wouldn’t use ‘fewer’ here, not because we’re not sure what the increment is, but because we’re measuring relative to a fixed point on a continuum, and whether something is to one side or the other of a point on a continuum is true whatever unit you happen to be using.
Yes, but not further than that, which is why it’s a discrete measure, and therefore why ‘fewer’ is more commonly thought of as correct. We count (discrete) items, but we measure (continuous) distance, temperature, etc.
Again, it’s not clear how you’re defining ‘benchmark’ here, and how that’s relevant.
I disagree, and I suspect that Mark Allen’s point is probably the correct one: that larger numbers, even if they remain discrete, are less jarring (rather than more correct) when used with ‘less’ rather than ‘fewer’. So ‘less than 100 words’ is less unacceptable than ‘less than 10 words’, and ‘less than a million words’ is less unacceptable still. Nothing to do with a ‘benchmark’ having been set, but something to do with the ability (or lack thereof) of our brains to grasp numbers. The larger they are, the more they begin to resemble blurry points on a continuum. This is probably especially true for people whose number skills aren’t as strong.
Closing: to be clear, the disagreement here mostly isn’t with which forms are acceptable or not, but why they’re acceptable or not. Your notion of a subtle pragmatic (rather than syntactic or even semantic) ‘benchmark’ not only flies in the face of conventional wisdom on the subject - which doesn’t make it wrong, but means that you have a case to make - but also hasn’t been well motivated here. The simple discrete/continous distinction (or the count/mass noun distinction, which is the same thing in other words) has far better explanatory power.
A quick addendum to my last comment, which hopefully will clarify things - and be a lot shorter.
Two quick examples, which have been used here already:
My claim, which I think is pretty uncontroversial, is that ‘less’ is preferred in [1] because temperature is continuous, and ‘fewer’ is preferred in [2] because number of socks is discrete. You count socks, but measure temperature. (Actually, I think we’d probably say ‘colder than 20 degrees’, or ‘below 20 degrees’, but never mind about that. “It’s less than 20 miles to New York” might be a clearer example.)
I believe you asserted that, in [1], a ‘benchmark’ has been set, which is at least part of the reason why ‘less’ is acceptable. But you also asserted, in reply to Pavel, that when counting socks, a benchmark isn’t set, so ‘fewer’ is okay. I think the implication is that were a benchmark set, ‘less’ would be preferred.
My question is, if you’re claiming that the discrete/continuous distinction isn’t relevant in [1] and [2], and that the crucial difference has to do with the setting of a benchmark or not, what is it about [1] which means a benchmark has been set, and about [2] which means that a benchmark hasn’t been set? What makes them different?
“Less” does work to qualify as well as quantify, whether we are talking about distinct objects or not. In math class, we may be asked to find a whole number less than 10, and we even have a “less than” symbol to guide us. I may have less of a haul from the store because, in simple terms, you have 12 items and I have nine. Now, rather than asking shoppers to make a judgment call, it seems that “10 items or fewer” is the better sign. I accept the argument for the sign that ends in “less.” I think it’s the lesser of the options, but that doesn’t mean it’s wrong.
Mark:
Just for what it’s worth, this example is entirely consistent with the discrete/continuous heuristic. Asking for a whole number less than 10 is a matter of finding a point on a continuum, relative to another point on the same continuum. We’re not counting things here; we’re dealing with abstract quantities on a continuous scale.
Again, this is consistent with the discrete/continuous heuristic. ‘Haul’ is a mass noun, rather than a count noun, and therefore continuous. The fact that in some specific cases a haul might be of discrete items (socks, say), and in other cases a mass of uncountable stuff (manure to fertilise the garden, say), is neither here nor there. If we choose to refer to it as a ‘haul’, we’re in the linguistic world of mass nouns, which are (linguistically) continuous.
We might say ‘a haul of 20 items [or fewer]‘, I suppose, in which case the language we’ve chosen to use is specific about it being a haul of discrete items. Here, our judgment is probably exactly the same as that for the supermarket checkout. It’s ‘items’ which determines whether we use ‘less’ or ‘fewer’, not ‘haul’.
As an editor, I agree with Paula and Pavel. Their explanations are consistent with how I’ve understood the distinction.
My goodness. I’m grateful to Paul (sorry about the “Paula”) for distilling his questions for me. I fear I’m not as succinct in my reply (I’m not known for being taciturn).
First, no, I’m not saying at all that the discrete/continuous distinction isn’t relevant. I think it is one of the reasons that we prefer less to fewer even when it would seem, on the surface, that we’re talking about countable things because the words in the expressions under consideration are plural (degrees, hours, and other similar increments we’ve mentioned). But I do think it is not a complete explanation.
Perhaps it’s the word “benchmark” that is troublesome, and I’m not wedded to the term. You can call it what you like—a level, a threshold, a spot, whatever. A given point on a continuum is a spot above or below which one can “go.” I am arguing that the requirement of 10 items is just such a spot, even though it does not conform to the idea of continuously break-downable increments that, it seems to me, is what you’re asserting to be the conceptual requirement for less rather than fewer to apply.
Note that in my original post I did not say that “ten items or fewer” was wrong. I said that “ten items or less” was not. I’m arguing that there is a good, solid, logical reason for less to be preferred in that context, which is that the people who made the policy set a limit, a spot above which one could not go.
If I understand your argument correctly, you’re saying that for less to be the most acceptable choice, two conditions must apply: that the stated amount must be a location on a continuum, and that the continuum must be infinitely granular—in other words, we could keep breaking down 24 hours from single hours to minutes to nanoseconds, which is why “24 hours or fewer” is awkward. But “10 items or fewer” is, apparently, not awkward because the continuum on which “10 items” sits is not reducible to partial items.
I think this is an arbitrary distinction. It’s the fact that an amount has been placed on a continuum at all—whether it’s a thermometer, a time clock, a pressure gauge, or the possible number of items a cart could hold—that is the key to the preference for less. The spot on the continuum is just as notionally singular as a rate is.
We can see this if we write out the two options to the controversial sign as complete sentences:
Your basket may contain only ten items or fewer items than that.
Your basket may contain only ten items or less than that amount.
In the first, “fewer” is not conceptually related to the limit of “ten items”; it’s doing its primary job of modifying the plural noun that follows it. The focus is on the items, not on the threshold.
In the second, “less” is conceptually related to the singular “amount,” which itself refers to “ten items” as a singular entity, as a threshold above which one cannot go.
It may well come down to which ending of the sentence one is supplying in one’s head, but my point is that the use of “or less” is fine because it reflects a very common pattern that is found in reference to stated benchmarks, levels, thresholds, spots…
I’m beginning to think that this issue is actually about mathematics, rather than linguistics. See below.
Wendalyn:
No, it’s not the word that’s the problem; it’s the lack of a clear, formal definition, whatever word you prefer to use. How can we tell whether, according to the concept you’re using, one of these ‘benchmark’ thingies has been set?
Okay, but it’s still not clear to me in which situations you feel this applies, and in which situations you feel it doesn’t. As I discussed in an earlier comment, you claim that the supermarket’s “ten items or less/fewer” does set a pragmatic ‘benchmark’ (or whatever word you prefer), but Pavel’s “fewer pairs of socks” doesn’t set a ‘benchmark’. Other than the discrete/continuous distinction, how are they different?
And, forgive me, I think you’re wrong about that. If it’s preferred by some/most people - and I’m actually not arguing that it might not be - I don’t believe it’s anything to do with a benchmark having been set, so much as simply the fact that for most people ‘less’ encompasses both discrete and continuous measures. ‘Fewer’, in contrast, is quite an esoteric word, as is - for most people - the difference between discrete and continuous. ‘Less/smaller than’ and ‘bigger/more than’ are enough to capture what they want to capture. I honestly don’t think your complex pragmatic basis for less being preferred by some people holds up at all. A simple collapse of ‘less’ into both meanings - smaller on a discrete scale, and/or smaller on a continuous scale - is enough to explain it.
I would add, in passing, that my personal preference is indeed for ‘10 items or fewer’, mostly, I think, because in my mind the difference between discrete and continuous measures is very clear, and using ‘less’ in that context violates my sense of which is which.
Yes - more or less - but your description here is illuminating. The two conditions you ascribe to me are actually one and the same condition. ‘Infinitely granular’ is what a continuum is. Time is a continuum, as are distance, temperature, weight, etc. The number of socks I happen to own, or the number of items I happen to have in my shopping basket, aren’t numbers on a continuum, exactly because they can’t have any arbitrary value.
This is what I meant by this possibly being a mathematical issue, rather than (only) a linguistic one. The distinction between discrete and continuous is mathematical, and I’m not sure it’s being fully taken into account here. The idea that ‘less’ is preferred for continuous values, and ‘fewer’ is preferred for discrete values, only makes sense if one has a good sense of what discrete is, and what continuous is, and what makes them different.
And here is an example of what I see as a crucial confusion. It seems that, for you, temperature, time, pressure, number of items, all exist in the same sense on a continuum. That suggests that your mental picture of the mathematical difference between discrete and continuous isn’t a clear one, or at least isn’t an important one. The number of any (countable) items is not on a continuum (in the sense that’s clearly relevant here), and is in a fundamental sense different from a measure on a continuous scale.
Your idea that the choice between ‘less’ and ‘fewer’ is something to do with whether it’s related to a singular or plural word is a red herring, made more convincing because it does often lead to a good choice, but for the wrong reason. The issue here isn’t singular/plural (or ‘benchmark’/no ‘benchmark’), but count (discrete) noun/mass (continuous) noun. What’s causing the confusion is that these often overlap. ‘Items’ is plural, and ‘amount’ is singular, but what actually determines a preference between ‘less’ and ‘fewer’ is that ‘items’ is a count (discrete) concept, and ‘amount’ is a mass (continuous) concept - or vague at best. “Ten items or less than that amount” doesn’t read all that well to me, but to the extent that it’s acceptable, that’s because ‘less’ is (as you say) being applied to ‘amount’, not to ‘items’, and (as you don’t say) because ‘amount’ - like ‘haul’ in a previous comment - is a mass (continuous) noun, not a count (discrete) noun, and not because it’s singular.
So your argument that “ten items or less than that amount” is (at least) more acceptable seems to me to be correct, but I don’t buy your explanation of why. I continue to argue as clearly as I can that it’s the simple discrete/continuous distinction that has the explanatory power here.
I really think that the idea of a continuum only gets you so far. The trouble with the mathematical concept is that a continuum is never really continuous, because you can always reduce it to ever more infinitesimal increments, so that in theory one can never go from point A to point B because there is an infinite number of smaller points to move between first.
In the real world, we do manage to go from point A to point B. We plot points. We set limits. And we say that one rate or amount is above that limit, and another is below it. And that limit, that point, is a singular concept—and by the way, mass nouns do not come into play because singularity is peculiar to “one.” (For example: “That’s one less thing to worry about.” Not “one fewer thing.”)
We don’t say “You may have twenty pairs of socks. No more, no less.” (Well, not unless we are so disturbed that we have more fundamental things to address than a more – less debate). But if we were to say it, the most idiomatic way to say it would be “no less.” Why? Because we’ve set a limit. A benchmark. A point above or below which one may or may not go.
And that, my friends, is the point beyond which I’m not going to go in this discussion, not because I want to shut down the dialogue, but because I have other work to do! Please feel free to continue the discussion if you like, and I’ll chime in if I really can’t resist doing so.
Wendalyn:
I’m not really sure where to go with the above. I do think on some fundamental level the difference between discrete and continuous quantities just isn’t being grasped. The idea of being able to reduce some measure to “ever more infinitesimal increments” isn’t something that makes it “never really continuous”. It’s exactly what makes it continuous. It’s why we can never know the exact value of pi, for example, and why we can never know exactly how far it is from place A to place B, or how tall a person is, or what the temperature is outside. All of these are approximations, to some accuracy/precision, because they all lie on true continuums. On the other hand, the number of socks I own, or the number of items in my shopping basket, can be unequivocally counted, and known precisely. These are two quite different concepts.
The idea of never being able to go from A to B is a bit of philosophical quibbling (it’s usually called Zeno’s paradox http://en.wikipedia.org/wiki/Zeno’s_paradoxes) that’s bizarrely irrelevant here.
That’s an interesting example. I agree that “one fewer thing to worry about” is awkward, but (at least for me) “one thing fewer to worry about” is fine, so (again, at least for me), I don’t think it’s the singular/plural issue, so much as an inability of “fewer” to go in certain syntactic locations. I suspect that you might not like “one thing fewer” for the same reason that you don’t like “ten items or fewer”, but both work well for me. Similarly -
“I have one sock fewer than yesterday.”
“I took an item out of my basket, so now I have one item fewer.”
- both work very well for me - almost certainly for the discrete/continuous reasons I’ve hammered into the ground. I’ll also point out that a quick corpus study (i.e., a Google search) brings up plenty of real-world uses of “one fewer thing”, so it’s not inaccessible to everyone.
Here’s a nice piece which discusses both the general less/fewer issue (and which is entirely consistent with my analysis), and also covers the “one fewer thing” wrinkle in some detail:
http://timtfj.wordpress.com/2009/06/29/fewer/
And some gratuitous links to other accounts I agree with, including the OED:
http://grammar.quickanddirtytips.com/less-versus-fewer.aspx
http://www.askoxford.com/asktheexperts/faq/aboutgrammar/lessfewer
http://www.gmathacks.com/sentence-correction/gmat-idiom-less-vs-fewer.html
Fowler’s Modern Usage is also entirely in line with the count (discrete) noun/mass (continuous) noun distinction.
Wow, after all these amazing comments I think rather than fewer (or less) questions, I actually have more! Good thing the word “more” works in opposition to both fewer and less. Too bad we don’t just have one word to go with “more.”
This issue falls into the “think I’ll rewrite the sentence instead of battling it” mode for me.
When my brain is not too tired, I think I will read through these comments again with less (fewer?) moments of aggravation!
Paul, since you’ve basically accused me of being thick (the passive “what’s not being grasped” may be a more polite wording, but it’s aimed at a single interlocutor), I’ll respond.
I’ll repeat: I do not claim that the discrete/continuous distinction is invalid, nor that it does not account for a majority of cases. I merely assert that it is inadequate to explain all cases.
My goal was to make the case that if you treat “10 items” as a notionally singular antecedent, as one does when one completes the implied sentence as “Ten items, or less than that amount,” then less is entirely appropriate. If instead the complete sentence in your head uses fewer as a premodifier (”Ten items, or fewer items than that”), then fewer will, as you put it, “work” for you.
In the link you posted to The other blog without a name, the writer (who also seems to be nameless; I cannot discover the name anywhere on the site) says of “Fifty words or less” that it “is interesting because, as Fowler points out, it is standard wording for English exams. A whole number of words is definitely what’s wanted. But the emphasis is still really on the length of the passage to be written, not on the individual words.”
Well, exactly. The emphasis is on the length of the passage, on the limit that was set. A notionally singular concept. In my view, that’s a benchmark, but it’s clear that we could not settle on an understanding of that idea and went off into mathematical red herrings, so I’ll leave that be.
The Cambridge Grammar of the English Language, written by a team led by two linguists who campaign against prescriptivism and who scrupulously adhered to descriptivist principles in the grammar (such that they only ever make assertions about frequency based on corpus analysis, rather than stating that a given form is “right” or “wrong”), say the following of constructions such as “You pass if you make ten mistakes or less/fewer” that have or after a numeral:
“less is the usual form here, with fewer quite marginal.”
Bottom line: if even the descriptivists’ descriptivists are willing to call the construction “quite marginal,” then nobody has any business correcting supermarket signs.
Wendalyn:
It probably wasn’t polite, no, but it also wasn’t an overstatement of my frustration at the mangling of the mathematical concepts here. When you say things like:
it’s clear that your internal model of what it means for something to be continuous isn’t consistent with the mathematical one - which is the relevant one to contrast with discrete in this context.
There are two issues here. The first is just the basic, accurate representation of the mathematical concepts. The second is the question of whether or not how we see the underlying mathematics affects the language we prefer when we represent those concepts. It’s simply the case that something being larger or smaller than something else (and in what way it’s larger or smaller) is a mathematical concept, and it must influence our language choices to some extent. To me, the difference between discrete and continuous is crucial. On the other hand, I strongly suspect it’s true that the difference between discrete and continuous values doesn’t occur at all to many/most people when choosing words. A corollary of that suspicion would be that if the discrete/continuous distinction doesn’t occur to them, the difference between ‘less’ and ‘fewer’ also wouldn’t occur to them.
More on this below.
In all cases for all people, then yes. But there are people out there - and I’m one of them - for whom the discrete/continuous distinction has extremely good explanatory power. I haven’t seen a single example in this discussion for which it doesn’t work that way for me, and my suspicion is that that derives not from (only) syntactic choices, but from a clear separation in my mind between discrete and continuous quantities.
But, yes, clearly there are many people out there who don’t follow that heuristic.
Fair enough. And that might well be what’s happening for you when you process things. But I don’t think you’ve convincingly argued for it as a process that’s happening generally.
My hunch - as I described in an earlier comment - is that the preference for ‘less’ in that situation (and others like it), is no more or less than the fact that most people are very happy to prefer ‘less’ over ‘fewer’ in all situations, irrespective of singular/plural or discrete/continuous criteria. ‘Less’ encompasses all manner of ’smaller than’ measures, just as ‘more’ encompasses all manner of ‘bigger than’ measures.
Why is that my hunch? Two reasons. First, it’s a simpler explanation - and it’s a good rule of thumb that we should look to simpler explanations first, until we know we need something more elaborate. Second, the preference for ‘less’, even in situations where the traditional discrete/continuous distinction would suggest that ‘fewer’ was more appropriate, doesn’t depend on the existence of the ‘benchmark’ you’ve proposed. ‘Less’ is preferred in far wider examples than that. Some quick real examples from the ‘net:
“Most Twitter users have less than ten followers.”
“With the default setting of ten, any move that has been played in less than ten games will be marked as a “no play” move for Fritz.
“The new committee had less than ten people who stood up to speak up for them.”
‘Less’ just semantically subsumes ‘fewer’ for most people, in almost all contexts.
Right, we don’t agree. I think there are at least three things other than your proposal going on here. The first is simply the above: that ‘less’, for most people and in most cases, can mean both ‘less’ and ‘fewer’. It’s not a matter of singular/plural; merely that ‘less’ doesn’t differentiate between those for most people.
The second is that ‘X words or less’ has become a kind of cultural catch-phrase, and its syntax isn’t really reconsidered any more. It comes whole, like a Lego brick of language, and persists for that reason.
And the third was discussed earlier: that the count/mass noun distinction can begin to blur with larger numbers, so even unequivocally discrete values, such as ‘fifty words’, can begin to feel like mass nouns. That’s mostly what the author was getting at, I think. When he said that “…the emphasis is still really on the length of the passage to be written, not on the individual words”, I read that as saying that the important thing is that the “length of the passage” is a mass noun, not that it’s singular.
That’s not an argument I necessarily disagree with - at least not very strongly. Certainly it’s a problem when people make a change because of perceived authority, rather than because it’s one they agree with.
There are two issues in all of this: one is which form is preferred, and the other is why it’s preferred. I don’t have any argument with the position that some people prefer one (or don’t care either way), and other people prefer the other. My problem in this case is mostly with your rationale for why those people who prefer ‘less’ do so. I think there are far simpler explanations generally, which have greater explanatory power.
Personally, I do have a preference for using ‘fewer’ in situations where the count is discrete. I like the idea that the difference between the concepts of discrete and continuous can be captured in language. The collapse of both concepts into a single word, ‘less’, is, in a small way, a loss of both precision and expressiveness. But I don’t doubt that’s the way things are going.