After class, a student came up with her grammar book, and wanted help working through some question tag exercises.
One sentence read:
It'll be a year before we see him again...The student was then prompted to chose between "won't it?" or "won't we?"
"The correct answer is won't it? " I told her.
"It's so confusing," she said. "I don't know which subject goes with the question tag. Isn't we the main subject?
"No, It is the main subject of this sentence," I said. "We is in a subordinate clause."
The very next sentence was:
I believe he's given up smoking... The choices were "hasn't he?" and "don't I?"
My native speaker intuition was telling me that "hasn't he?" sounded much more natural in this sentence. But "I believe" was clearly the main clause of the sentence, so this was contradicting what I had told her before.
Update:
I looked this up in Practical English Usage by Michael Swan. Although Swan does not explain why, he does note that this is the case.
As Swan says:
Note the use of question tags in sentences beginning with I (don't) think and similar expressions
I think he's Norwegian, isn't he? (NOT...
(Swan p.472)
But is there a reason for this?
This comment has been removed by a blog administrator.
ReplyDeleteStephan, sorry I accidentally deleted your comment when I hit the wrong button. I can't seem to be able to get it back, but I have the contents saved from my e-mail notifications, so I'll just repost it below. Sorry about that.
ReplyDeleteHi again, Joel! ^_^
I was browsing your 'unanswered grammar questions' series, and I decided to give this one a shot, since it's relatively straightforward.
So, there's a few things going on here. First, I'd say that in all cases, with all other things being equal, the tag can refer back to something either in the main clause or in a subordinate clause. In the think-and-similar-expressions case, the reason the tag has to refer back to something in the subordinate clause isn't grammatical -- it's psychological. That is, practically speaking, it's normally the case that we have access to our own thoughts, so it would be unusual for someone to question whether or not they believe something. But we can invent scenarios where this kind of tag actually becomes quite normal:
Imagine you're directing a play, and in the scene that's being rehearsed, one of the actresses is trying to come up with a natural reaction to her husband coming home smelling like cigarette smoke. She says to you "I believe he's given up smoking, don't I?"
Here, the tag makes sense, because the person asking the question doesn't have direct access to their own thoughts (or, more precisely, the thoughts of the character they're playing). So it doesn't really look like there's a grammatical prohibition on this, but the context definitely matters.
As for why a tag can't refer back to something inside a 'before' clause: this one's grammatical, but it doesn't necessarily have anything to do with syntax (more like semantics, or pragmatics). Most (perhaps all) subordinating conjunctions are presuppositional, which means they take for granted as true whatever clause happens to follow them. So, "It'll be a year before we see him again" takes for granted that we'll see him again. We can tell that this really is a case of presupposition, because when we try to negate the sentence, the assumption lingers; "It won't be a year before we see him again" still seems to mean that we'll be seeing him again (perhaps within a year). And since a presupposition is information that the speaker assumes he or she can take for granted as being known, it would be bizarre to go on and question that assumption.
What's going wrong in that first sentence of yours is that same thing that's going wrong in a sentence like "John's children are awfully noisy, but does he have children?" It's strange to question something that you've already decided to take for granted as true.
Stephan, thanks a ton for that explanation. I read it over, and it all makes perfect sense to me now. (Although I can't believe these questions were popping up in an ESL grammar book. This is some pretty heavy stuff!)
ReplyDeleteYou got me thinking about this, and I dug a little deeper; lo and behold, I discovered something more (and ended up learning something new)!
ReplyDeleteSo, I think the first half of my response still works -- about being able to question either the main clause (in the right context, anyway) or the embedded clause (at least with think-verbs). But it seems the second half isn't quite right: while it's true that it's generally strange to question a presupposition, as in the "John's children" example, that can't be what's going on here.
If it were true that we can't ever question a presupposition, then we shouldn't be able to say something like "I know he's given up smoking, hasn't he?" To the degree that it's acceptable (I think it works in a context where you say it like you're sort of questioning what you thought you knew about someone, or you're unsure of your own memories about them), it's a problem for my explanation. "Know" is usually considered to be a factive verb, which means it presupposes the truth of the clause that comes after it. This is easy to tell with a positive/negative pair of sentences like "Mary knows he's given up smoking" and "Mary doesn't know he's given up smoking"; each one assumes it's been given up. But if "know" is factive, and we can also question the embedded clause, we're sort of back to square one.
It seems, instead, that you were right all along! The accepted pattern is what you first thought it was: tag questions must refer back to main clauses, not embedded ones. Which of course leaves us wondering why it ends up working with verbs like "believe," "think," and even "know." Well, apparently, these are exceptions! But not random ones, thankfully. The relevant distinction is that these are so-called 'bridge verbs' (as opposed to non-bridge verbs). So, what this means is that we can say things like "What do you believe he bought" or "What do you think he bought." You can't really do this with non-bridge verbs like "whisper"; "What did you whisper he bought" sounds pretty strange. Sure enough, tag questions referring back to bridge verb clauses work, while tag questions referring back to non-bridge verb clauses don't! Compare the good "I think he's Norwegian, isn't he?" with the bad "I whispered he's Norwegian, isn't he?"
So, there you go! At least, until I find another problem with this explanation. ;-P
Interesting. I didn't even realize the problem in your first explanation, but I guess this gets more and more complicated.
ReplyDeleteWhat about your original point that we can invent scenarios where the question tag would be normal, like the actress in the play saying "I believe he's given up smoking, don't I?"
I think that still holds up. Besides fiddling with the context, Another way to get around the apparent awkwardness of "I believe he's given up smoking . . . don't I?" is to just switch the subject, so that it no longer refers to the speaker: "You believe he's given up smoking . . . don't you?" works just fine. In fact, when the subject is "you," the tag question referring back to the main clause (rather than the embedded one) seems much more natural -- to me, at least.
ReplyDeleteTo make things even more complicated, I'm not sure the whole bridge/non-bridge explanation actually, totally works -- at least, not without some kind of qualification. Both "think" and "regret" are bridge verbs, since both of the following questions seem to sound okay -- to my ears, anyway: "Who do you think never got the chance to speak" and "Who do you regret never got the chance to speak." And yet, "think" is happy with a tag question poking around inside its embedded clause, while "regret" seems not to like it very much: "I think he got the chance . . . didn't he?" is fine, but "I regret he got the chance . . . didn't he?" seems odd (though, maybe you can think of a context that'll make it work).
So, the full pattern seems to be that tag questions target main clauses, except when certain verbs are in play (e.g., think, believe, suppose, figure, guess) -- and even then, only when the subjects of those verbs refer to the speaker. Otherwise, the tag still prefers to refer back to the main clause. So, it's a very narrow set of exceptions. But, what exactly these verbs have in common with each other is tricky to pin down: it doesn't look like it's bridge verbs, as one grammarian suggested, since it doesn't work with "regret"; it doesn't look like it's only non-factive verbs, since it seems to work with "know" (in a very specific context, mind you); maybe it's some complex combination of properties? Like, maybe this works with positive emotive attitude verbs -- so, verbs like the ones I listed just above -- but excludes negative emotive words like "regret", "surprised," and "ashamed," which allow so-called negative polarity items like "any" and "ever" into the sentence. It might also have something to do with whether a verb is 'epistemic' in nature (i.e., having to do with knowledge and belief).
Yikes!
It's obviously a very complex pattern, and it's no wonder it's so hard for learners to get a grip on it. It also looks like it's not something that's been written about much, which is why it's been so hard for me to come up with a definitive, well-informed answer. One published paper* speculated this is because it's something that's common in spoken speech, but not so common in writing. I'll definitely keep thinking about it, and I'll try to talk to some experts about it, to see what I can find out. For sure, we'll definitely be doing more episodes about different kinds of verbs; I can't promise I'll write any scripts about tag questions and embedded clauses anytime soon, though, since there seems to be very little written about them! :-/
*I kid you not, it's called "English tag questions are quite complicated, aren't they?"
Wow, that is complicated, isn't it?
ReplyDeleteIt makes you really pity the poor ESL student.
I've often heard people say that English is so ridiculously complicated that it should never have become the common global language. This seems to confirm that sentiment.
But maybe all languages have some sort of weird complications?
I'd wager every language has this kind of stuff floating around in it, based on my experience. My master's thesis was all about an even more complex pattern in Quebec French. These kinds of things aren't really obvious to native speakers, until they're forced to try to spell them out, either to themselves, or to someone else.
ReplyDeleteAnyway, I hope I didn't overwhelm you with all my thinking out loud! It's just what tends to happen when you give a linguist a language puzzle. ;-p
No worries, I was confused about question tags from the get-go, so it couldn't have gotten much worse. And I was tracking with you through the explanation.
ReplyDeleteWhen I teach ESL, it always amazes me how complicated grammar is, and how much my brain does without me even being consciously aware of it.
I remember thinking the same thing a couple years ago when teaching my students when to omit the relative pronoun.
http://joelswagman.blogspot.com/2014/01/omitting-relative-pronoun-worksheet.html