Most questions, when converted into statements, retain their overall "meaning", i.e. the statement is asserting what the question is asking.

Question: Can you grate the pears?

Assertion: You can grate the pears.

Yet, I recently came upon the following sentence (it happened to come out of my own mouth) that doesn't follow this pattern:

Question: Can you play anything by Kapustin?

Assertion: You can play anything by Kapustin.

To me, these sentences ask and assert different things. The first asks if there exists a thing by Kapustin which you can play. The second asserts that for all things by Kapustin, you can play them.

And yet, the noun phrase "anything by Kapustin" didn't change, nor, I would argue, did its grammatical context in the sentence. But somehow "anything" takes on a different meaning when in a question vs. a statement. It doesn't have to be the object, either - look to the sentence "Can anything happen?" vs. "Anything can happen".

Are there more of these nouns? A complete characterization of them, or a name for them? Is there some underlying grammatical explanation for their existence or is this just a quirk of colloquial English?

  • Your second assertion can also mean exactly the same thing as the second question: Q: What you would like me to play? A: You can play anything by Kapustin. But there are many words, phrases, and sentences that can have multiple interpretations, so it's not clear to me how to get a specific answer out of this. Aug 1, 2019 at 21:06
  • I think you are blaming anything for the sins of can. Compare with can you play something by Kapustin. In a question, can can mean would you please or are you able to.
    – Phil Sweet
    Aug 1, 2019 at 23:08
  • This is not really about English.
    – Lambie
    Apr 23 at 21:10
  • Have you considered that part of the meaning of many sentences is the context in which they are uttered? What you seem to be doing is to utter the sentence with no context and then find a context in which the contextless sentence could be taken as false. There are in fact a higher proportion of words whose meaning can shift, given different contexts, than we imagine. Many (possibly most) grammatical 'rules', definitions and 'classifications', especially in English, have more potential nuance of this kind than we realise.
    – Tuffy
    Apr 23 at 21:27

1 Answer 1


First, let's take these two sentences:

  1. You can play anything by Kapustin.
  2. You can't play anything by Kasputin.

One would assume that, when we replace can by can't, it should negate the sentence, so that (2) is false whenever (1) is true and vice-versa. But that is not the case: if you can play some things by Kapustin but not all of them, then (1) and (2) are both false (on the most obvious interpretation of (2)).

This is because there are two senses of the word any (and, correspondingly, two senses of the word anything). (My source here is Huddleston & Pullum (2002), pp. 382-383.)

The first of these is the free choice "any," which essentially means that, when some arbitrary member is chosen from a set, the statement is true of it. So (1) means "If we choose an arbitrary work x by Kapustin, you can play x."

The second of these is the "non-affirmative" any. This can only occur in sentences that aren't affirmative statements, e.g. in negated statements like (2). It's essentially a counterpart of some: when we negate "I said something," we get "I didn't say anything." So "You can't play anything by Kapustin" is the negated version, not of "You can play anything by Kapustin," but of "You can play something by Kapustin." (The free choice "any" can also occur in non-affirmative contexts, but it's not the usual interpretation.)

Now let's consider a third example:

  1. Can you play anything by Kapustin?

Importantly: yes-or-no questions ("closed interrogatives") generally count as non-affirmative contexts, so the non-affirmative any can be used there (ibid., p. 832). In that context, some is also valid: both "Did Kim make some mistakes?" and "Did Kim make any mistakes?" are correct, and have the same meaning, though the version with "some" suggests that the answer is more likely to be "yes."

So, in "Can you play anything by Kapustin?" we are likely to interpret "anything," not as the free choice anything found in (1), but as the non-affirmative anything found in the most likely interpretation of (2). So the question isn't asking "Is it true that, for any work x by Kapustin, you can play x?" but rather "Does there exist some work x by Kapustin such that you can play x?"

  • 1
    But the meaning changes in the dialog "What should I play?" "You can play anything by Kapustin." Nov 25, 2023 at 21:09
  • @PeterShor Indeed. That's the difference between the "dynamic" and "deontic" use of can; can can mean either "be able to" or "be permitted to."
    – alphabet
    Nov 25, 2023 at 21:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.