A grammar puzzle

English grammar can be strange. Sometimes it seems to have a rule of breaking its own rules, as it were. An example that occurred to me yesterday involves the words less and fewer. I’ll tell you about it shortly, but I need some background first.

A rule

First consider the basic rule about these words:

• fewer refers to things you can count
• less refers to things you can’t count: continuous quantities.

So, for example, if you drop one of your plates on the floor while washing up and it shatters, you have fewer plates than you did before. But if you eat a larger piece of cake than planned, there is less cake left than there would have been otherwise.

Some exceptions which aren’t

This seems like a clear enough rule, and it’s one which we mostly adhere to in written English. Fowler [1], though, mentions a few apparent exceptions, such as

• It is less than seventy miles to London
• [It] costs less than fifty pounds
• We have had reliable temperature records for less than 150 years
• [Please write] fifty words or less.

On closer inspection, the first three of these turn out to fit the rule.Try using fewer in those examples: It is fewer than seventy miles to London sounds (at least to me) as though the distance to London has to be a whole number of miles. But it isn’t: distance is a continuous quantity, which just happens to be measured in miles. We aren’t counting the number of miles, but measuring the distance.

Similarly the time since temperature records began is unlikely to be a whole number of years, and we’re measuring the time, not counting individual years; less than 150 years really means “A time whose length is shorter than 150 years”.

The money example seems slightly different, since money does come in distinct steps. However, the steps are pence, not pounds. Normally you won’t get the correct price of something by counting out a number of pounds, and (at least to my ears) fewer than fifty pounds sounds like doing precisely that. We generally think of money as a thing which we have a lot or a little of, not as a pile of coins which we have many or few of. So it still fits the rule.

Fifty words or less 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.

As a borderline case, Fowler  gives having had in his house at one time no less than five Nobel Prize winners. I’m less happy about that one: I think that in written English it should definitely be no fewer than five of them. Nobel Prize winners seem to me to be something that you definitely have a whole number of, not something that you measure out. On the other hand, maybe when you have a crowd of Nobel Prize winners your attention is on the size of the crowd rather than on the individuals. But I somehow doubt it. Nevertheless, these examples do in fact fit the rule: fewer for things you count, and less for things you don’t. I’m merely a little dubious about the idea of not counting Nobel Prize winners.

Speaking colloquially

I’ve not studied the speech aspect of this, but it’s clear that many (maybe most) people ofetn don’t adhere to this rule when speaking. They say things like I’ve got far less things to do today and might treat less things to do, less work to do and less to do as equivalents which are all variants on the idea of doing less.

Are they speaking “incorrectly”, or are they using a different set of rules of “correctness” for speech? I’m not sure: to me this one does feel more like not noticing a word that doesn’t fit properly, thereby getting it wrong,  than like using a differnent rule  to determine what fits. But at the same time, carefully using fewer can feel artificial at times, creating too much formality, especially in situations where it’s harder to say or when speaking to someone who doesn’t use it. Far fewer feels more awkward to the mouth than far less, for example, and I think there probably is a “rule” in speech  of using phrases which have a smoother sound to them. Maybe sounding nice sometimes takes precedence over “correctness”. But whether it actually sounds nice to the listener will of course depend on how alert they are to the grammatical structures, how bothered they are by it, and whether it affects clarity of meaning.

The puzzle

OK, that was rather a lot of background. Here is the puzzle, though it may be that I just hear things a particular way which other people don’t share. I’d like to hear other people’s opinions on it. It concerns the situation where someone, say, has some things to  do and then does one of them. The situation afterwards can be expressed in a variety of ways. Some feel more natural than others; some seem more grammatically logical than others. This is how it looks to me, though it may be different for someone else:

Acceptable

• I now have fewer things to do.
• I now have one less thing to do.
• I now have one thing less to do.

These all feel to me like natural ways of saying it.

Less acceptable

• I now have less things to do.
• I now have one thing fewer to do.

Less things to do feels natural as a spoken expression, but either wrong or borderline as a written one. One thing fewer is verging on awkward: not exactly wrong, but not a very natural expression either.

Unacceptable (to my ears)

• I now have one fewer thing to do. (No no no! One less!)
• I now have one fewer things to do. (“One things?!”)

And there’s the puzzle.

• How many things do you have to do? Fewer. You have  fewer things to do.
• How many fewer? One. You have one fewer things to do.

And yet, far from being the correct expression, one fewer things to do is the most unambiguously wrong one of the lot.

On the other hand, if two tasks are done instead of one,  fewer becomes OK again: I have two fewer things to do.

At first sight, this is all very puzzling. The questions are ones like these:

• Why can I have fewer things to do but not one fewer things to do?
• Why can I have  one less thing to do but not one fewer thing to do?
• Given that the things to do are ticked off my list one by one, why does less rather than fewer end up being the apparently correct word?

Attempt at an answer

The problem with the two versions I listed as “unacceptable” seems to be  that however much we may want them to be logical, they refuse to read that way.  One fewer things to do insists on reading as though fewer qualifies one things, so singular and plural are mismatched. One fewer thing to do tries to make fewer thing into a valid element. But we know that fewer applies to more than one of something. What’s rather strange, though, is that one less thing to do doesn’t seem to suffer from the same problem, even though less thing isn’t really any more valid an element than fewer thing. Maybe it’s chosen simply because it doesn’t leap out quite so blatantly.

(By the way, I really wish I had a quick way to draw some sentence diagrams here.)

From my list, I think the three “least incorrect” versions are

• I now have one less thing to do.
• I now have one thing less to do.
• I now have one thing fewer to do.

In all cases, there is one thing which has already been done. This leaves less/fewer [things] to  do. But if we test the structure by removing fewer or less from the sentences, we see that they are all versions of I now have one thing to do, which isn’t the situation. The one thing is precisely the one which doesn’t need doing.

So the grammar is confusing because it’s trying to have its cake and eat it. Structurally, less to do and fewer to do are firmly attached to the one thing. But in their actual meaning, they refer to something entirely different: all the other things to do, which are nowhere to be seen in any of the three sentences. They’re trying to refer to two conflicting things at once.

Finally, if we remove the one thing from the three sentences and see what is left, we end up with

• I now have less to do
• I now have fewer [things] to do.

If the most natural phrase were I now have one thing less to do, this would give us our answer: it means “I now have less to do, by one thing”. Interpreted this way, the sentence is grammatically correct and self-contained. However, I think the most natural one is one less thing, not one thing less, so either some illogicality remains, or something unidentified is still going on.

I would welcome any thoughts on this! In particular, on whether my feeling as to the relative acceptability of the different constructions coincides with yours. There may be regional or international differences, or you might simply have gravitated towards a different usage from mine.

Note

[1] R W Burchfield (ed.). The New Fowler’s Modern English Usage, Oxford University Press, 1996, p. 295. Back

Books I ought to finish reading

Just for fun, here’s a list of them. As it happens, they’re also books I want to finish reading but keep forgetting to, or doing something else instead. In no partcular order (actually, the order in the pile):

Books to finish

• Miles Kington, How Shall I Tell The Dog?
• Oliver Sacks, Musicophilia: tales of music and the brain
• Stephen Fry, The Book of General Ignorance
• Stephen Fry, The Book of Animal Ignorance
• Steven Mithen, The Singing Neanderthals: the origins of music, language, mind and body
• Robin Dunbar, The Trouble with Science
• Seth Lloyd, Programming the Universe: a quantum computer scientist takes on the cosmos
• John D Barrow, Impossibility: the limits of science and the science of limits
• Rodney Huddleston, English Grammar: an outline
• Barry Green, The Inner Game of Music
• Andrew George (trans.), The Epic of Gilgamesh
• Eknath Eswaran (trans.), The Upanishads
• Stephen Fry, Stephen Fry’s Incomplete and Utter History of Music
• Roger McGough, Collected Poems

Some of those are books I’ve started, some I’m half way through, some I’ve nearly finished . . . and maybe some aren’t exactly for finishing, since really they’re for dipping into.

Actually, one of the most interesting of those is also one of the most demanding to read: the grammar book. It’s not, as you might imagine, a guide on how to write; it’s a very concentrated analysis of how English grammar works, and I see that on the next page I have a section which starts

Constructions involving a non-finite as complement of the predicator exhibit a great deal of diversity and complexity; they present formidable problems for the analyst—and it is not surprising that widely varying accounts are to be found in the literature. One problem is this. The prototypical complement is an NP, which is why we speak of the occurrence of non-finites in complement function as involving nominalisation.

All of which does in fact make sense, but it’s not the kind of material that effortlessly goes into the brain, especially if it’s a few months since you were last reading the book and need to remind yourself what a predicator is and what is or isn’t being nominalised, i.e. being treated like a noun. Let’s just say that once we start looking at how English grammar actually works, it makes languages like German with nice, rigid, clearly-defined rules start to look a lot more straightforward than English.

Maybe I’ll focus instead on the Miles Kington book, which has stuff like this coming up (see, I can’t help reading ahead):

Dear Gill,

People are making a lot of money out of self-help books these days, and I would like you to be one of those people.

By helping to promote my new self-help book.

Which would be about self-pity.

Did you notice in my first letter that I referred to the jumble of self-pitying thoughts I first had when I was diagnosed with cancer?

My immediate response was to be apologetic for this stance, because we are always taught not to be sorry for ourselves, as if there were something dreadfully feeble about it. There are no nice words in English at all for ’self-pity’. There are lots of disapproving ones. Whingeing, sulking, moping, etc., etc.

(Personally, I think we are entitled to indulge in a little self-pity when we are told we have cancer, as long as we disguise it as something else. Shock, a nervous breakdown, long sobbing fits. Something like that.)

But self-pity is so common that it earns no respect at all, only disapproval, as in phrases like: ‘Sitting around all day feeling sorry for herself,’ or ‘You’d think he was the only one who had ever had leukaemia.’ Which quickly leads to phrases like: ‘Why doesn’t she just pull herself together?’ and ‘Cheer up dear—it’s only bi-polar disorder!’

My brilliant idea would be to turn it all round and treat self-pity as a potentially positive force.

This certainly seems to be a brilliant book, from the 40% or so that I’ve read in its intended order. Miles Kington wrote it in the last months of his life, when he knew that he did in fact have cancer and might well die from it. It takes the form of supposed letters to his literary agent about ideas for books he might write about the situation, but is really a humorous but heartfelt look at attitudes encountered and so on. Very entertaining, but also thought-provoking.

But that’s just one list of books. Here’s another:

Books to start

The main reason I haven’t started the books in this list is that I don’t have them. They’ve been recommended, or mentioned, by other people:

• Paul Davies, About Time
• [I don't know the author], The Universe is a Green Dragon
• Peter Bernstein, Against the Gods: the remarkable story of risk
• Daniel M Wegner, The Illusion of Conscious Will

Now that’s a much shorter list, but I’ve a nasty feeling that’s simply because of having forgotten to make a note of them all . . . Oh dear. I wonder what’s missing . . .

To unashamedly split infinitives

“Superstitions”

Warning: You probably need to be a language enthusiast to enjoy this article. But I’m hoping you are one!

I expect you’ve heard of Fowler. I mean H W Fowler’s Modern English Usage, the standard reference book for grammar questions, first published in 1926. Besides being authoritative, it was quite entertainingly written. The edition I own is the 1996 one, edited by R W Burchfield.

And here is an entertaining entry I happen to passionately agree with:

superstitions. Among the most enduring of the superstitions or myths about our language are these: sentences should not begin with and or but; sentences should not end with a preposition; and infinitives should not be ’split’. For further examples of such beliefs, see FETISHES.

A certain kind of person makes a virtue out of picking up such “errors”. If they hear one on the radio, or read it in a piece of serious writing, they will protest–maybe even to the extent of writing to Feedback or the publisher. They will take a delight in “correcting” the “offender”. (Ironically, publishers will avoid using such a person as a proofreader or editor: an important skill is the ability to know when not to change things. Even more ironically, the person will usually object to the use of hopefully in the way in which I’ve just used ironically, as a “sentence adverb”.)

They wouldn’t like the sentence I used just now to introduce the quotation. It started with and, split the infinitive to agree, and ended with the preposition with, thereby breaking all three “rules”.

But to be honest, I don’t like happen passionately to agree. It feels clumsy; I can’t “happen passionately”; it feels awkward breaking up happen to by sticking passionately in the middle of it; splitting passionately agree by inserting to feels uncomfortable too. The rules they want me to use when writing aren’t the ones my brain uses when reading. Here, then, is an entertaining entry with which I happen passionately to agree” . . . That’s simply too formal and pedantic in style for a piece of writing like this. It’s also longer, because of having to include which.

Take another example. Some people insist that none is always singular, so that you must never say none of them are, only none of them is. (This is the same, I think, as assuming that none can only mean not one, never not any.) Fowler says this:

none. It is a mistake to suppose that the pronoun is singular only and must at all costs be followed by singular verbs or pronouns. It should be borne in mind that none is not a shortening of no one but is the regular descendent of OE [1] nan ‘none, not one’. At all times since the reign of King Alfred the choice of plural or singular in the accompanying verbs, etc. has been governed by the surrounding words or by the nominal sense.

I’ll go with King Alfred, then.

Good grammar or rigid grammar?

What is “correct” grammar for?

Let’s ask the more basic question: what is language for? It’s for communication. English is good when it clearly communicates what it means to, and bad when it doesn’t. Communication is helped if words mean the same thing to the speaker as to the listener, or to the writer as the reader; similarly it’s helpful to have grammatical constructions which unambiguously mean the same to both parties. Good grammar is grammar which achieves that.

Most linguisticians [2] these days see grammatical rules not as prescriptions for correctness, imposed by some authority, but as descriptions how the language normally works. They exist because conventions have evolved about what the different constructions mean. These have become more or less “standardised”. Why are they there? To enable clear communication. How did they arise? They evolved that way, through usage. They didn’t arise from someone deciding what was or was not “correct”, but from experience of what worked and what didn’t. Language is a continuous experiment.

Rules for their own sake are also, it seems to me, rather a blunt instrument. A nice feature of English is its ability to express many fine shades of meaning. I think this comes partly from its huge vocabulary, and partly from its flexible word order. That’s one of the things I most love about it.

So I get upset when people mechanically impose rules like these; they actually lose the fine distinctions. To my ears the supposedly incorrect phrase often has a slightly different meaning from the “correct” one. The actual “correct version” is the one which most accurately represents the intended meaning.

A small audience

Here’s an ecample. I was startled by an acquaintance’s responce to my question: “Has it ever occurred to you that It was a small audience and They were a small audience mean completely different things?” “No,” she said, “Audience is singular, so it can only be It was . They were is wrong. I was married to an English teacher for twenty-five years, so I know these things.”

My protestations that Fowler didn’t agree got me nowhere. And she didn’t see the mental image that amused me, either . . .

Can you see it? Let me add a few words:

1. It was a very small audience: there were only five people in it.
2. They were a very small audience: not one of them was over a foot tall.

Clearly [3] a small audience of a few humans is more likely than a small audience of hundreds of tiny humanoids, but it was small and they were small mean quite distinct things.

You might still be feeling uncomfortable about this. Actually there is a slight problem with the second sentence, but not the one she was complaining about. In one case it’s the audience which is small, and in the other it’s they who are small, but I use small audience both times. And so the two situations are actually

1. The audience was very small.
2. The audience were very small.

Shades of meaning

Here is a more fun example, involving a split infinitive. I’m not sure why this is the sentence which came to mind when I first thought about this, but it was! Consider the rebuke:

• You don’t need to suddenly start jumping out at people.

Our friendly pedant would say that to suddenly start is wrong, and that it should be suddenly to start. But I think this either changes the meaning or makes it ambiguous, as you’ll see in a moment.

Let’s be thorough about this. Let’s try putting suddenly in all places where the sentence will let it go. Just for the hell of it. Let’s see what happens.

1. Suddenly you don’t need to start jumping out at people.
2. You suddenly don’t need to start jumping out at people.
3. You don’t suddenly need to start jumping out at people.
4. You don’t need suddenly to start jumping out at people.
5. You don’t need to suddenly start jumping out at people.
6. You don’t need to start suddenly jumping out at people.
7. You don’t need to start jumping out suddenly at people.
8. You don’t need to start jumping out at people suddenly.

I’ve not included jumping suddenly out because it seems to me to stretch things further than they will really go, and I’ve not included at suddenly people because it’s too ungrammatical to make any sense.

Each placing gives a different meaning, a different range of meanings, or a slightly different emphasis. Some of the differences quite subtle and hard to describe, but I’ll have a go.

1. Suddenly you don’t need to start jumping out at people.
2. You suddenly don’t need to start jumping out at people.

These both mean almost the same thing, but with different emphasis. In (1) the situation suddenly changes into one in which starting to jump out at people is no longer required. In (2), you suddenly change, into a person who no longer feels the need to start jumping out at people.

3. You don’t suddenly need to start jumping out at people.
4. You don’t need suddenly to start jumping out at people.
5. You don’t need to suddenly start jumping out at people.

The first and last of these are unambiguous. In (3), the sentence is disagreeing with your claim that although you once didn’t need to start jumping out at people, now you suddenly do. It refers to a sudden change in need. In (5), the infinitive-splitting version, you weren’t jumping out at people in the past, but have suddenly begun to do so. The reference this time is to a sudden change of behaviour.

(4) though, need suddenly to start, is ambiguous, even though it is the supposedly correct version, which our friendly pedant would want. We can read it as “need suddenly” or “suddenly start”. I myself find it easier to read as “need suddenly”, so the meaning of the “corrected” version is different from that of the original.

6. You don’t need to start suddenly jumping out at people.
7. You don’t need to start jumping out suddenly at people.
8. You don’t need to start jumping out at people suddenly.

At first sight, these all mean the same thing: you weren’t jumping out at people; you are just starting, or about to start, to behave that way; you don’t need to make that change.

But if I weigh them carefully, each seems to me to have a different emphasis. As I hear it, (6) seems either to treat suddenly jumping out as a single unit, complaining about the action as a whole, or to emphasise jumping out more than suddenly. In (7), on the other hand, jumping out suddenly seems to emphasise suddenly as well, and there are now two complaints: “You don’t need to start jumping out at people, and what’s even worse, doing it suddenly so they all have heart attacks”.

(8), jumping out at people suddenly, emphasises suddenly even more, and the sentence is now definitely ambiguous. It can either be the same kind of meaning I’ve just described for jumping out suddenly at people, or a new one which is approximately “I know you like jumping out at people, and we’re all used to it, but there’s no need to start being sudden about it”. Presumably you previously issued some kind of warning first, or did it in a very predictable way.

In each case, we’re talking about sudden jumping. The most natural place for suddenly is immediately before jumping. Putting it after instead of before emphasises it; putting it even later emphasises it even more.

What split infinitives are for

Back to our split infinitive: to suddenly start. What does it do? It acts as a sort of “container”. Anything which is between to and start can only refer to start. It’s a remarkably strong container: however far away we move the to, the result is unambiguous. In fact it’s even possible to carefully, so as to visibly make my point, place one split infinitive inside another. The preceding sentence did just that.

On the other hand, look what happened when we moved suddenly . . . start apart:

• don’t need to suddenly start: unambiguous
• don’t need suddenly to start: awkward and now ambiguous
• don’t suddenly need to start: unambiguous but a different meaning
• suddenly don’t need to start: yet another different meaning.

The adverb suddenly seems to attach itself to any likely-looking verb in the vicinity, producing lots of different meanings. The split infinitive keeps it under control, unambiguously attached to start.

Interestingly the author of a book I have on computer programming seems to have adopted a policy of always splitting infinitives. Once I noticed, I couldn’t find a single instance where he’d avoided it. I suspect this was because he wanted to avoid any hint of ambiguity at all, since programming needs complete precision.

So I’m unhappy about the split infinitive “rule”. I think Fowler was right to describe it as a “superstition”. I think a much more valid rule would be: don’t separate an adverb (e.g. suddenly) from its verb (e.g. start) any further than is absolutely necessary.

Why does the “rule” persist? I suspect it might be because

• once you’ve learnt it, it’s an easy rule to remember
• in any instance, it’s totally clear whether the rule has been applied or not
• people learnt it at school when they were young, and simply accepted what they were taught
• finding mistakes in written English is fun.

In my opinion the split infinitive “rule” is cultural, not linguistic, and it’s one we could well do without. I wish English teachers would refrain from teaching it to their pupils. It’s more important that they can recognise and avoid ambiguity, and the “rule” is a hindrance to that. Ambiguity and clarity is actually more interesting as well, since there’s lots of scope for fun examples of sentences which have gone wrong. The kinds of sentence, in fact, which gets sent in to The News Quiz or Have I Got News For You.

Notes

[1] Old English. Back
[2] people who study linguistics–as distinct from linguists, who learn languages. Someone who speaks eight languages fluently is a linguist; someone who is an expert on the way children learn to talk, or the way dialects evolve over time, is a linguistician. Back
[3] Another “sentence adverb”, as Fowler calls them.Back