Why are certain there-sentences infelicitous in English?



The Cambridge Grammar of the English Language states that the first three of the following four excerpts are semantically or pragmatically anomalous (to give that term some context, it cites We frightened the cheese as an example of an anomaly):

  • President Clinton appeared at the podium accompanied by three senators and Margaret Thatcher. Behind him there stood the senators.
  • President Clinton appeared at the podium accompanied by three senators and Margaret Thatcher. Behind him there were the senators.
  • President Clinton appeared at the podium accompanied by three senators and Margaret Thatcher. Behind him there was the Vice President.
  • President Clinton appeared at the podium accompanied by three senators and Margaret Thatcher. Behind him there stood the Vice President.

It goes to explain that the existential (be) characteristically requires that a definite displaced subject (the senators, the Vice President) be addressee-new and the presentational (stand) occurs more readily with a discourse-new displaced subject.

  • Are the authors of CGEL correct that the first three examples are pragmatically anomalous in major English dialects?
  • Why are the existential and presentational constructions in there-sentences sensitive to addressee-status and discourse-status of the postverbal noun phrase, respectively?

EDIT: Apparently, the sensitivity to definiteness of NP is called definiteness effect, and Google Scholar kindly provides tons of papers about it. However, all that linguistick-y stuff is sending my brain into meltdown mode, so it would be great if someone summarised it as an answer, if those papers happen to answer the question.

EDIT2: As per a request by @Cerberus and two other people who have upvoted his comment, the existential here refers to the there be construction, the presentational has some other verb than be (e.g. There remain many problems), the addressee is the one to whom something is addressed, addressee-new stands for something new to the addressee, discourse-new stands for something new to the discourse, addressee-status may be either addressee-old or addressee-new, discourse-status may be either discourse-old or discourse-new, and a displaced subject is “the phrase that corresponds to the subject of the syntactically more basic construction” (A Student's Introduction to English Grammar by Huddleston, Pullum).


Posted 2011-09-22T13:18:10.810

Reputation: 268

5Interesting. Your question is packed with technical terms that you might want to explain, like addressee, presentational, and discourse status.Cerberus 2011-09-22T13:26:02.363

"Unacceptable"? An interesting question, and certainly one of the reasons a forum like this exists but I think they're all perfectly understandable to a typical English speaker, which may be what actually matters. – None – 2011-09-22T13:38:18.013

@mickey: The CGEL considers it bad style; but I wouldn't be surprised to see any of those sentences written by a native speaker on a website of mediocre quality. Pullum (major author of CGEL) is usually against style advice, unless he considers something pretentious language (my conjecture).Cerberus 2011-09-22T13:44:00.813

4@mickeyf I suppose that depends on whether you are interested in how English language works, or can be satisfied with stringing together words, grunts, and gestures to try to get your point across.Kit Z. Fox 2011-09-22T13:44:01.673

2I don’t understand. If the first, ‘Behind him there stood the senators’, is semantically or pragmatically anomalous, why isn’t the fourth, ‘Behind him there stood the Vice President’? Moreover, while I can see that ‘We frightened the cheese’ is semantically anomalous, I don’t quite see what bearing it has on the other four sentences.Barrie England 2011-09-22T14:27:42.663

1@BarrieEngland According to CGEL, “the Vice President” is discourse-new and so would be an acceptable noun phrase in the presentational construction. On the other hand, “the senators” have already been mentioned in the discourse previously (“… accompanied by three senators”) and therefore that information is discourse-old.Vitaly 2011-09-22T14:35:28.547

3I agree that "Behind him there were/stood the senators." are unacceptable. The others are ok. I can't hope to answer the next part. Btw, I think it helps to just repeat the whole sentences instead of leaving <...> there. – None – 2011-09-22T14:41:29.097

1@Vitaly Now that you mention it I did understood the senators in first and second sentence as different beings from the ones, who walked with Clinton to podium. But still I don't understand why the third one, introducing a new being the Vice President, is anomalous. It sounds more natural to me than there stood construction. – None – 2011-09-22T14:42:04.453

@Vitaly. Thanks.Barrie England 2011-09-22T14:43:48.733

@Philoto “The Vice President” is discourse-new (that is, appears in that particular piece of discourse for the first time) but addressee-old (the hearer would know the Vice President). The existential (there be) doesn't accept addressee-old subjects, according to CGEL.Vitaly 2011-09-22T14:58:23.173

@Vitaly very odd. I must admit I have difficulty finding an eligible counterexample to this rule, but still Behind them there was Vice President sounds more natural. – None – 2011-09-22T15:25:16.367

@Philoto: Do you mean to put the in there?Daniel 2011-09-22T17:01:49.427

@drɱ65 δ Yep, I meant to put it there, my mistake. – None – 2011-09-23T06:14:34.267

You mean to say that CGEL would not approve of "there was the moon", since it's never addressee-new? That phrase is used often enough it gets charted by google Ngrams. – None – 2011-09-23T11:04:31.967

@PeterShor — Firstly, “And then there was the Moon Maiden,” “Then there was the moon cult,” “There was the Moon Tree Planting Ceremony” (those are real examples from Google Books) are valid Google results for “there was the moon.” Secondly, Google Ngrams doesn't quite show “the moon” in all those matches to be addressee-old. Thirdly, CGEL lists a few cases when such usage is possible, viz. addressee-old entities treated as addressee-new, addressee-new tokens of addressee-old types, addressee-old entities newly instantiating a variable, etc, etc.Vitaly 2011-09-23T11:48:55.803

I can quote results from Google books, too: "I went outside the house and there was the moon, just a thin crescent with Venus bright near it." "But when she came to the foot of the old staircase there was the moon shining down from some window high up"; "and that all of a sudden the sky cleared, just before sunrise — and there was the moon." "I made the proper correction on the chart and, sure enough, there was the moon right where it ought to be." None of these seem odd to me at all. – None – 2011-09-23T12:00:59.587

@PeterShor — And fourthly, language patterns aren't quite as strict as mathematical ones. If the majority of observed usage patterns (not even all of them) comply with some “rule,” it can be considered sufficient.Vitaly 2011-09-23T12:01:24.050

@Vitaly: I think the difference between these examples and the CGEL examples is that "there was" in the CGEL examples is a static description, whereas the "there was" in the moon examples is more dynamic -- all of a sudden the moon appears to the speaker, when it hadn't been visible before. If the CGEL isn't including this case, then they've missed part of the usage of "there be" in English. – None – 2011-09-23T12:09:34.317

@PeterShor — It seems to be that in your first two examples the Moon is treated as an addressee-new entity; and I think that in the last example “the Moon” does not refer to the Moon but rather to some representation of the Moon in the chart. I am not sure about the third example, though. Anyway, I am editing the body of my question to include the words “characteristically” and “occurs more readily” to make it closer to CGEL's wording. Thanks for pointing that out.Vitaly 2011-09-23T12:19:04.787

1+2: This should be illegal. I upvoted twice!? Once before migration, once after. Ah well.Daniel 2011-09-24T23:15:16.220



There are roughly two kinds of accounts of the definiteness effect.

  • One, as suggested by CGEL, is that the "pivot" (DP following the verb) must be discourse-new, or novel, or something along those lines. (this idea is due to McNally, Comorovski, Zucchi, and others, just in case CGEL omitted the citations, as it often does.)
  • Another, which has been somewhat more prominent than the explanation CGEL points to, is that it is some logical property of the determiner that plays the key role. (Keenan, Barwise and Cooper, Higginbotham, and others. But especially Keenan.) I suppose this is the linguisticky stuff that is causing melting. But Keenan's proposal in particular has a tremendous amount of empirical coverage, in a way far more than the discourse novelty approaches. If you want a good recent summary of the Det approaches you should check out chapter 6 of Itamar Francez's 2007 Stanford dissertation, I think available on his website. Most of these accounts amount to realizing the intuition that the constraint ensures that to check the truth of an existential claim, all you need is the set of individuals semantically provided by the "coda" (the thing optionally following the pivot), and not any larger set.

There is no good account of what unifies these two types of explanations (though maybe McNally's dissertation takes the best stab), or really even why either of them should be true. The best idea I'm aware of, which as yet has only been pursued informally is sketched in ch. 6 of Francez's dissertation (due to David Beaver), and is roughly that the non-canonical word order in existentials marks that the pivot must be topical, which leads to the definiteness effect. How to turn this into a predictive theory, I haven't a clue (though it is something I'm working on).

In general Francez's dissertation is a good starting point without having to read through every google scholar result.


Posted 2011-09-22T13:18:10.810

Reputation: 529


I think CGEL is pretty much correct here.

It's at least "odd" to use there in this construction to say something additional about the people already mentioned in the previous sentence, so that makes the first two versions sound somewhat clumsy to me.

I also don't like the third version, but I'm not sure CGEL are on such solid ground with their obscure addressee/discourse-new terminology to explain why it's not so good. If I replace "behind him there was the Vice President" with "in front of him there was the Presidential seal" it doesn't seem so awkward to me. The fact that the VP is another politician/person similar to the other four already mentioned seems to be significant to my inner ear, which winces a bit at #3.

Although I can't justify it on grammatical grounds, I don't much like #4 either. I agree "stood" is better than "was", but if the VP isn't mentioned in the first sentence that implies he was already on the podium. Which makes it seem strange to suddenly add him into the verbal picture that's being painted by using "appeared" in the first place. The context implies we already have a mental picture of the podium before the presidential appearance, so bringing up that pre-existing detail as if it were part of the sudden apparition being described just seems clumsy to me.

In short, #4 may be grammatical by CGEL's lights, but I'm not keen on the entire construction. I think that's because "stood" doesn't really mean anything different to "was" here. We're at the margins of fine distinctions, but I'd rather just sidestep the issue with "The VP stood behind him".


Posted 2011-09-22T13:18:10.810

Reputation: 522


I think envisioning these sentences without the "behind him" included makes it clear why the first three examples are incorrect.

For example:

President Clinton appeared at the podium accompanied by three senators and Margaret Thatcher. There were the senators.

Here the function of the word "there" is clearly to introduce a new person/object into the context. Adding a spatial reference ("behind him") doesn't really change the purpose of that sentence, in my mind; it just adds extra information.

Just as the example I've written is redundant (in that we already knew the senators were present), so too is the original example.


Posted 2011-09-22T13:18:10.810


1What about the veep? – None – 2011-09-23T10:38:14.530

Oops! Good call. I didn't read the third example carefully. I think the idea of introducing new information could still be argued for, though. For instance, 'Behind the president, there was a statue dedicated to lost sailors' is fine, but 'Behind the president, there was the statue dedicated to lost sailors' is off again. Perhaps because the VP exists as an entity anyway, such that we can use a direct article, he doesn't qualify as a 'new' piece of information. – None – 2011-09-23T13:01:02.523


Now that I’ve seen the full sentences, I think the answer to the first part of your question, at least, is, in principle, yes. To avoid the loaded ‘unacceptable’, however, I would say instead that the first three are unlikely to appear in any serious piece of prose written in Standard English.

Barrie England

Posted 2011-09-22T13:18:10.810

Reputation: 276