The English-Learning and Languages Review Æ Homepage


A critical examination of the account of

 some and any

(and their compounds)

in Quirk et al.’s

A Comprehensive Grammar of

the English Language (1985)

(The false trail of grammatical categorising: 2)



Amorey Gethin


For my account of the use of some and any, see




Incorrect description of the difference between some and any

Their rigid ‘assertive’/’non-assertive’ scheme

“Apart from negative contexts, non-assertive items appear in a number of other contexts.”


‘restrictive relative clauses’

‘morphologically negative and negative import words

Quirk et al.’s biased redefining of sentences

‘basically negative and non-assertive sentences’

False correlation of ‘assertive’ forms with expecting answer “yes”

Misunderstanding the negative question

More misunderstanding of conditional clauses

Quirk et al.’s basic ‘assertive’/’non-assertive’ confusion

Their superficial but tortuously frustrating analysis; conclusion




Incorrect description of the difference between some and any


Near the beginning of Quirk et al.’s account of some and any they do attempt a differentiation of the words in some sort of meaning terms.


The primary difference between some and any (and between the some- and any- compounds) is that some is specific, though unspecified, while any is nonspecific. That is, some implies an amount or number that is known to the speaker. This difference tends to correlate with the difference between positive and negative contexts:

I have some money on me. [a specific, though unspecified amount of money]

I don't have any money on me. [an unspecified, and also nonspecific amount of money; no limit on the amount is assumed] (pp.783-84)


Here straight away we find an example of how impossible it is to define words properly with other words. Some very often does not imply an amount or number known to the speaker. For instance:


I’ve got some money somewhere, but I don’t know how much, and I’ve forgotten where I’ve put it.

I’m sure there are some people who would disagree with you.


Those are examples where clearly the amount or number is unknown to the speaker as well as to their interlocutor. But even in less obvious cases the quantity is likely to be unknown:


There are some books in the other room.


Indeed, an important motive for using the word some is very often precisely that speakers cannot specify the amount, even if they want to.


The description in square brackets of any is illogical. The whole point here is that I haven't any money, so there isn't an amount of money to be either unspecified, nonspecific or unlimited! Some might regard this as hair-splitting, because, they might say, any is a quantifier, and it is as a quantifier that it needs to be defined. But any is not in fact a quantifier, as Quirk et al.’s own example demonstrates. With one exception, it is never absolutely necessary to use any with a plural, although more often than not we do. That is almost certainly because when we use any we are more often than not thinking in plural contexts. The exception, of course, is when we use the phrase any of. But any in any of the books is no more a quantifier than best is a quantifier in the best of the books. Some complete sentences with any may seem to have a quantifying aspect, such as


Have we any milk?


but others, such as


If you think of any solution, let me know.


do not at all.


Their rigid ‘assertive’/’non-assertive’  scheme


Some, Quirk et al. say, is ‘assertive’, while any is ‘non-assertive’. They are led to this conclusion, I suspect, because any is so often found in negative sentences, and so they repeat the old error of giving a basic negative character to the word. I think it must at least in part be the error they have made in other places in their grammar, that is to say, confusing the meaning of individual words with the meaning of whole sentences. A good case can be made for saying that any, far from being a non-assertive word, is an extremely assertive word. That, surely, is the way native English-speakers experience it, though perhaps a better way of describing any would be to call it an emphatic word. Any has a much more ‘assertive’ or emphatic feel to it than some, quite apart from the fact that it means something completely different.


Anyone with any sense would have stayed at home.


However, Quirk et al. use the vague attributes of ‘assertive’ and ‘non-assertive’ to claim that the use of some and any is determined by the grammatical category of their context. Some, they say, is found in ‘positive’ sentences, any in negative, interrogative and conditional sentences.


They appear to ignore a fundamental principle concerning words. Virtually every word has its own unique meaning. In some sentences – i.e. in certain combinations of meanings - this means that it cannot be replaced by a different word without radically changing the meaning of the sentence. But in other combinations of meanings, replacement by an alternative word does not effectively change the meaning of the sentence. Some and any are themselves good examples of this principle. There is little or no practical difference in meaning between


Do we have some milk?


Do we have any milk?


But the radical difference in meaning of


I didn’t understand some of his jokes [I didn’t understand a certain proportion of his jokes]


I didn’t understand any of his jokes [I understood none of his jokes]


is obvious. They could have made everything clear and simple from the start by emphasizing this essential difference in meanings. Instead, they subject the problem – and they make it into a problem where there actually isn’t one – to a confusing mixture of abstract semantic and abstract grammatical analysis.


On page 784 they say:


“Apart from negative contexts, non-assertive items appear in a number of other contexts.”


 They include:

(i) Yes-no questions that expect a negative response or are neutral in expectation (11.6f)

Do you know any of the teachers here?


One is not told what "appear" means. Does it mean "can appear" or "must appear"? If "can appear", the information is useless, as one still does not know when and why one uses any rather than some. If "must appear", it is clearly wrong, as the very first of their examples illustrates. (i) could just as well be, with little effective change in meaning in this particular case, and still keeping neutral expectation:


Do you know some of the teachers here?




Quirk et al. continue their list of 'non-assertive' items in other than negative contexts, from which the following are a selection:


(iv) conditional clauses (15.33ff)

If anyone ever says that, pretend not to hear.


But one can just as well say:


If someone ever says that, pretend not to hear.


‘restrictive relative clauses’


(vi) restrictive relative clauses modifying generic noun phrases, where the clauses have conditional meaning:

Students who have any complaints should raise their hands. ['If students have any complaints, they should raise their hands.']


If some students are able to follow and even apply that complicated analysis, they should be made aware that one can also say


Students who have some complaints should raise their hands.


with a slight, though not very significant, change in meaning.


‘morphologically negative and negative import words’


(vii) after words that are morphologically negative or that have negative import:


You still have time before you have any need to register. ['You don't have any need to register now.']

He's too old to play any rigorous games. ['He doesn't play any rigorous games.']

They can prevent any demonstration.


As regards the first of these three examples under (vii), we would have to say, for instance,


It takes time before you can master some finer points of grammar. [There are certain finer points of grammar you won’t be able to master yet.]


if that is what we mean. Any would give a different meaning. In the same way, the second and third examples could also have any replaced by some, with a radical effect on the meaning.


He's too old to play some rigorous games. [But he can play certain other rigorous games.]

They can prevent some demonstrations, but not others.


(It is perhaps worth noting here, although the point is not immediately relevant, that some in They can prevent some demonstrations is not necessarily stressed. For instance: They can prevent some small local demonstrations, but in the face of a general strike they would be powerless.)


Quirk et al.’s biased redefining of sentences


But what is perhaps as significant as these elementary mistakes is how biased the paraphrasing is. The first one under (vii) changes the meaning considerably, and as for the second, it does not follow at all from the original sentence that he actually doesn't play any rigorous games. Perhaps he ignores his doctor's advice.


‘basically negative and non-assertive sentences’


Quirk et al.'s bias is even clearer in one of the earlier versions of their work, A grammar of contemporary English, (1972, Longman, pp.223-24):


Although the main 'superficial' markers of non-assertion are negative, interrogative, and conditional clauses, it is the 'deep', basic meaning of the whole sentence which ultimately conditions the choice of the some or the any series.

See further 7.44. For example, in the sentence


Freud probably contributed more than anyone to the understanding of dreams.


the basic meaning is negative and non-assertive, as appears in the paraphrase

Nobody contributed more to the understanding of dreams than Freud.


Note first the mistake again of confusing the meaning of the sentence with the meaning of the word (some/any). Even more fundamental is the apparent belief that the use of particular words is determined by some ‘deep’ grammatical category. But also, why should we only re-define in the direction assertive to non-assertive. Why not say that the second sentence (non-assertive) is basically the first (assertive)? And the argument is conveniently selective. I imagine Quirk et al. would not be eager to maintain, in regard to some, that


Bill can learn some things faster than others.

is ‘deeply’, basically

Bill cannot learn some things as fast as others.


Only some of the guests had ties.

is ‘deeply’, basically

Ties were not worn by some of the guests.


since their argument would dictate that any rather than some should be used in these sentences.


False correlation of ‘assertive’ forms with expecting answer “yes”


On page 808 of A comprehensive grammar of the English language we find


Like negative statements, yes-no questions may contain nonassertive forms such as any and ever (cf.10.60). The question containing such forms is generally neutral, with no bias in expectation towards a positive or negative response.

Someone called last night. Did anyone call last night?


But questions may be CONDUCIVE, ie they may indicate that the speaker is predisposed to the kind of answer he has wanted or expected. Thus a positive question may be presented in a form which is biased towards a positive answer. It has positive orientation, for example, if it uses assertive forms rather than the usual nonassertive forms:

(1) Did someone call last night? [ ‘Is it true that someone called last night?’]


(2) Do you live somewhere near Dover?


Once more they are selective as well as muddled. For instance, unlike (1),


Does someone know Italian?


is not, I think, biased towards a positive answer. (2) is a good example of how decisive particular meanings are. It is the fact that the question is about living near Dover (or any other specified place) that gives the question ‘positive orientation’. One expects “yes” just as much if one asks:


Do you live anywhere near Dover by any chance?


But the difference between somewhere and anywhere does indeed give the two questions different meanings, even if in this case again the difference is probably insignificant in practice. If the somewhere question has a different sort of context, it does not in any way expect “yes” rather than “no”. There is no difference in expectation between


Have you got somewhere to put this ice cream?


Have you got anywhere to put this ice cream?


Misunderstanding the negative question

Following a complicated analysis on pages 808-09, Quirk et al. write on page 809:


If a negative question has assertive items, it is biased towards positive orientation.

Didn’t someone call last night?


In other words they say that the someone indicates that the speaker expects the answer “yes”. Yet in a note on the very next page (p.810) they give the example (replacing their ‘assertive’ someone with their ‘non-assertive’ anyone)


Doesn’t anyone know the answer? [‘Surely someone knows the answer.’]


These appear to be mutually contradictory statements. They are muddled at least partly because they have missed the point that it is the negative question that expects “yes”, as in both 


Didn’t anyone call last night?


Doesn’t someone know the answer?


irrespective of whether anyone or someone is used. These last two sentences are further examples of how the difference in meaning between some(one) and any(one) sometimes has no practical significance. But the difference is still there as always. It is this difference in meaning, fundamental and often of crucial practical significance, the difference that really matters, that Quirk et al. effectively ignore in their pursuit of misleading irrelevancies.


More misunderstanding of conditional clauses

Their failure to illustrate that basic difference persists throughout the book. On page 1092 one finds:


Conditional clauses….are like questions in that questions are generally either neutral in their expectations of an answer or biased towards a negative response. Therefore, like questions, they tend to admit nonassertive forms.


If you had ever listened to any of my lectures, you would have known the answer.

She’s taking a stick with her in case she has any trouble on the way.


In their first example the meanings of ever and any fit well together. (It is a good example of what an emphatic, ‘assertive’ word any is.) But take away ever in the first sentence, and it is perfectly normal to replace any with some – and of course change the meaning in the process! (Some could be used instead of any in the second sentence without changing the practical significance.)


Quirk et al.’s basic ‘assertive’/’non-assertive’ confusion

In fact Quirk et al. start out on the wrong track much earlier in their grammar. In a note on page 84 they write:


Whereas it is frequently impossible for a positive statement to contain nonassertive forms (*I have any ideas) it is by no means unusual for assertive forms to occur in questions and negative clauses: Do(n’t) you have some ideas? Our use of the term ‘nonassertive territory’ does not exclude, and indeed anticipates, a more delicate stage of analysis….at which we acknowledge that assertive forms can give an assertive ‘bias’ to constructions which are predominantly nonassertive.


This typifies Quirk et al.’s whole approach, the complicated and wholly unnecessary chopping up into detailed specification and reservations and conditions. Most usage flows naturally from the basic meanings, not from such arbitrary rules prescribing a multitude of various usages. Their use of *I have any ideas to illustrate their claim that one cannot use ‘nonassertive’ forms in ‘positive’ statements is in fact a perfect demonstration of how it is really particular meanings that determine usage:


I welcome any ideas


is a perfectly normal sentence.


Their superficial but tortuously frustrating analysis; conclusion

The authors did not claim that their work would be useful for students of English as a foreign language. But I suspect that any who do venture into this Quirkian analytical morass would probably find it too frustratingly difficult to get anything out of - even if it had been correct.


Quirk et al.’s grammar is now 25 years old. But to judge from the books for students of English as a foreign language that I can find on the market, virtually nothing has changed. These books’ treatment of some and any usually just looks like an abridged version of Quirk et al. Most say that  some is more common in affirmative sentences, and any more common in negative sentences and questions. This is undoubtedly true. But once more it is only true because any situations, for instance, more commonly fit negative situations in LIFE. Simply saying that some (or any) is more common in a certain category of sentence is vague and unhelpful for non-English-speaking students. It tells them nothing about the distinction in meaning, which is the real, essential and only important distinction.


Minor revision and revised layout 2 May 2010


The English-Learning and Languages Review Æ Homepage