More people than you think will understand

« previous post | next post »

Andrew Dowd sends me a genuine, attested case of the kind of sentence that I have elsewhere called plausible angloid gibberish. It is a particular kind of mangled comparative that somehow seems English when it isn't. It has absolutely no right to be called grammatical, and nothing can explain why it is that we (falsely) believe that it has a meaning that could be accounted for in the regular way — it doesn't and it couldn't. No syntacticians that I know of can say why it sort of slips by, in comprehension and sometimes (as here) even in production. The sentence came from http://www.backspace.com/notes/2004/06/, citing AdAge, and it reads thus:

In Michigan and Minnesota, more people found Mr Bush's ads negative than they did Mr Kerry's.

Complete and utter syntactic nonsense. And yet when you read it you see what they meant long before you realize that they couldn't have meant it.

The meaning you wrongly pick up is something like "In Michigan and Minnesota, more people found Mr Bush's ads negative than found Mr Kerry's ads negative." But why? It doesn't say that! Andrew says, "It took me days to figure out there was anything wrong with it."

There really is a lot we don't know about syntactic processing.



42 Comments

  1. Kyle said,

    December 27, 2009 @ 10:40 am

    Is "In Michigan and Minnesota, more people found Mr Bush's ads negative than found Mr Kerry's ads negative" not the meaning? The "did" functions as a pro-verb and the "ads negative" is implied?

  2. axon said,

    December 27, 2009 @ 11:15 am

    @kyle:

    who is "they" ?

  3. msH said,

    December 27, 2009 @ 11:20 am

    Is it the interpretation of "they" that is strange? I suppose "they" means "people", not necessarily the same people as mentioned before, although they might be. It's a fairly vague "they" – it feels to me like singular "they", in the sense that there's something there we don't know and it doesn't matter. Singular "they" means we don't care about the gender, and this "they" means "the same or other people, we're not saying which". Why is that wrong? Are you going to explain?

  4. Bob said,

    December 27, 2009 @ 11:21 am

    The "more people" makes it sound like we're talking about two different groups' reactions: some find Bush's ads negative, others find Kerry's ads negative. But "they" makes it sound like we're talking about the same group having two different reactions: All people in general found Bush's ads to be more negative. I think the intended reading is the former, but my initial interpretation was the latter. I first read it as meaning: "In Michigan and Minnesota, people found Mr Bush's ads more negative than they did Mr. Kerry's."

  5. Simon Spero said,

    December 27, 2009 @ 11:23 am

    It's attested, it has a immediately ascertained meaning, and it took a professional linguist days to figure out that there was anything wrong with it.

    That's sufficient for provisional grammatical asylum under the European Charter of Construction Rights.

  6. Mr Punch said,

    December 27, 2009 @ 11:42 am

    This, I think, is another case of a sentence made less clear by an attempt to render it grammatically correct. It's that "they did," surely an unnautural addition, that causes the real trouble — "more people found Bush's ads negative than Kerry's" isn't great, but it's less problematic.

  7. James said,

    December 27, 2009 @ 12:03 pm

    @ Mr Punch: agreed.

    Here's a hunch. The writer had two thoughts. One of them is the one we are now focusing on. The other is the thought that people found Mr. Bush's ads more negative than they found Mr. Kerry's. (A closely related thought, but plainly not the same one: the comparative in this second one is between degrees of negativity, not numbers of people.)

    Now,

    People found Mr. Bush's ads more negative than they did Mr. Kerry's.

    gets the same verdict from me that

    more people found Bush's ads negative than Kerry's

    got from Mr. Punch, which is to say, not great, but acceptable. So maybe the problem came in the author's losing track of which thought was which.

  8. Mr Fnortner said,

    December 27, 2009 @ 12:08 pm

    Bob's got it right (never mind the role of "did" as a putative pro-verb). Is it a confused mind or an inarticulate tongue (pen) that produces such a sentence? It's common enough for speakers to start one sentence and finish another, but for writers?

  9. Dave M said,

    December 27, 2009 @ 12:12 pm

    I heard something similar on ABC, who used to claim that "more people get their news from ABC than from any other source." And it does indeed sound funny once you think about it. Still, what Simon Spero said (although I would remind him that the US is *still* not a signatory to the Charter).

  10. TL said,

    December 27, 2009 @ 12:37 pm

    In cases like this, I would greatly appreciate if Language Log also point out how to rephrase the sentence. I sometimes have to write similar comparative sentences, and often wonder what is the right way to do it.

    "In Michigan and Minnesota, more people found Mr Bush's ads negative than found Mr Kerry's ads negative." sounds a little bit too wordy for me. Like another commenter, I think the below sentence works: "In Michigan and Minnesota, more people found Mr Bush's ads negative than Mr Kerry's." Can someone please explain the objection to it?

  11. Franz Bebop said,

    December 27, 2009 @ 12:50 pm

    Maybe the writer is trying to avoid ellipsis, and just doesn't correctly identify what words were left out. It seems like the writer wants to insert a subject-verb pair, since that superficially "sounds correct" or complete, but in fact the true missing words are something more fragmentary. In fact, the elliptical sentence is a lot more clear than even the correct non-elliptical version.

    Both of your examples include the word "more". The pattern in the current example seems to be "more people {verb phrase} than {verb phrase}." No missing subject here, so inserting one causes trouble.

    The earlier pattern is "more people than just me {verb phrase}." I regard this as clumsy but not ungrammatical. "More" seems to be a clumsy synonym for "other." It might be impossible to write a complete (non-elliptical) sentence around this usage. The example sentence actually cited was "more people {verb phrase} than I {verb phrase}," which doesn't make any sense. Fortunately, there's a long list of potential replacement phrases ("other people besides me…", "I'm not the only one…"), so it's easy to avoid the problem.

    – More people have been to Russia than I have. (original)
    – More people have been to Russia than just me. (preserves the ellipsis)
    – More people than just me have been to Russia. (same thing but better word order)
    – I'm not the only one to have been to Russia. (avoids "more")
    – It's not just me. Other people have been to Russia. (avoids "more")

    – In Michigan and Minnesota, more people found Mr Bush's ads negative than they did Mr Kerry's. (original)
    – In Michigan and Minnesota, more people found Mr Bush's ads negative than Mr Kerry's. (preserves the ellipsis)
    – In Michigan and Minnesota, more people found Mr Bush's ads negative than found Mr Kerry's ads negative. (avoids ellipsis correctly, but is wordier and less clear)

    Just a wild guess: Maybe the writer is clumsily trying to avoid a garden path problem. When you start out one of these sentences, it's not necessarily clear which one you are going to end up with:

    1) More A {verb phrase} than B {verb phrase}
    2) More A {verb phrase} than {verb phrase}

    Maybe English speakers prefer pattern (1) and are less comfortable with pattern (2).

  12. Franz Bebop said,

    December 27, 2009 @ 12:58 pm

    @Dave: What's wrong with ABC's slogan? It grammatical and the meaning is clear. I see no problems with it.

  13. Shimon Edelman said,

    December 27, 2009 @ 1:31 pm

    @Geoff: "There really is a lot we don't know about syntactic processing." — I guess the same could be said about Santa Claus, and for the same reason.

  14. Athanassios Protopapas said,

    December 27, 2009 @ 1:32 pm

    Perhaps grammaticality has been overrated.

  15. Byron said,

    December 27, 2009 @ 1:39 pm

    FYI, the "plausible angloid gibberish" has been discussed in formal contexts. They discuss the classic examples of the type "More people have been to Russia than I have", and their analysis might extend to this example (though I'm not sure). In any case, I thought it should be brought up.
    Wellwood, A., R. Pancheva, V. Hacquard, S.W. Fults, and C. Phillips. 2009. The role of event comparison in comparative illusions. Poster presented at the 22nd Annual CUNY Conference on Human Sentence Processing.

    Phillips, C., M. Wagers, and E. Lau. To appear. Grammatical Illusions and Selective Fallibility in Real­Time Language Comprehension. Language & Linguistics Compass.

  16. John Thayer Jensen said,

    December 27, 2009 @ 1:45 pm

    I confess it made my head spin! Does it mean:

    "There were more people who found Mr Bush's ads negatives than there were who found Mr Kerry's negative."

    ??

  17. Ryan Denzer-King said,

    December 27, 2009 @ 3:27 pm

    I think there's a problem with saying "it doesn't mean that". What we mean when we say something is "wrong" or "impossible" for syntactic theory means we have the theory wrong. I gather that's pretty much what you meant with your post, but I think a lot of linguists fall into the trap of throwing away such utterances as performance errors or the like, whereas I think more likely these are real sentences uttered by competent speakers, and can and must be accounted for by a complete theory of syntax and semantics.

  18. Tom Recht said,

    December 27, 2009 @ 3:51 pm

    Could it be a kind of mongrel sentence, cross-bred from the following two (which are both grammatical, differ from it minimally, and have the intended meaning)?

    In Michigan and Minnesota, more people found Mr Bush's ads negative than did Mr Kerry's.

    X

    In Michigan and Minnesota, people found Mr Bush's ads negative more than they did Mr Kerry's.

    =
    In Michigan and Minnesota, more people found Mr Bush's ads negative than they did Mr Kerry's.

  19. Spell Me Jeff said,

    December 27, 2009 @ 4:18 pm

    @Ryan Denzer-King

    Did you miss the import of "There really is a lot we don't know about syntactic processing"?

    Cf the archived post to which we are linked: ". . . notice that the issue as I see it is not about whether the example type is ungrammatical, or whether it is merely semantically incoherent, or why. The puzzle is about why people initially don't notice there is anything wrong with it at all."

    Pullum's on your side.

  20. Peter Metcalfe said,

    December 27, 2009 @ 4:26 pm

    I think the writer actually omitted a word or two, depending on how grammatical you want the original sentence. Hence:

    In Michagan and Minnesota, more people found Mr Bush's ads negative than they did [for] Mr Kerry's [ads].

  21. Peter Metcalfe said,

    December 27, 2009 @ 4:28 pm

    Or substitute [with] for [for] in the suggested sentence.

  22. Andrew Johnson said,

    December 27, 2009 @ 4:55 pm

    What about this kind of thought process:

    They produce this correct utterance in their head: "more people found Mr Bush's ads negative than Mr Kerry's"

    But then style or prescriptivism or what have you kicks in and they second-guess themself at a higher (but perhaps still subconscious) level, thinking that the ellipsis in this particular case sounds stilted or pedantic. The the un-ellipsed version would be: "more people found Mr Bush's ads negative than [people found] Mr Kerry's [ads negative]"

    However, they don't realize that the initial 'people' and the second [people] are different groups (an unexpected semantic quandary), so they turn [people found] into 'they did' and voilà.

    Even if they had been thinking "…than *other* people found Mr Kerry's," "other people found" could be legitimately simplified as "they did", with the same results.

    It's like a counterbleeding order in phonology. They'd already processed the sentence for syntactic/semantic integrity, but then attempting to clarify the ellipsis after the fact torpedoed that.

  23. Mark Anderson said,

    December 27, 2009 @ 5:09 pm

    “In Michigan and Minnesota, more people found Mr Bush's ads negative than they did Mr Kerry's.”

    In a survey of N people, the number of people who found Mr Bush’s ads negative is Bn, the number of people who found Mr Bush’s ads positive is Bp, the number of people who found Mr Kerry’s ads negative is Kn and the number of people who found Mr Kerry’s ads positive is Kp. The number of people who did not express a view on Mr Bush’s ads or found them neither negative nor positive is Bq and the number of people who did not express a view on Mr Kerry’s ads or found them neither negative nor positive is Kq. According to the above statement, which of the following must be true?
    1. A single survey was conducted in Michigan and Minnesota
    2. Separate surveys were conducted in Michigan and Minnesota
    3. The results of the surveys were the same in Michigan and Minnesota
    4. B + K = N
    5. Bp + Bn + Bq = N
    6. Kp + Kn + Kq = N
    7. Bn > Kn
    8. Kp > Bp
    9. A majority of the people surveyed (ie >N/2) both (a) were Bn and (b) were not Kn
    10. Kn is a subset of Bn
    11. Insufficient information has been given to know whether any of the above statements is true.

    Students are advised not to spend more than 10 minutes on this question as their brains may explode.

  24. Andrew Dowd said,

    December 27, 2009 @ 5:40 pm

    If you want my theory on why this ungrammaticality is so easy to ignore, it has to do with the way we process variable binding in comparatives in general. Remember that comparatives more or less require ellipsis to be well-formed, so we're used to looking for a gap we have to interpret.

    So first you find the gap in the second clause of the comparative (after 'did'), and fill it in with whatever seems relevant (find X's ads negative) until you have a semantically complete clause. After that is done, you look back to the first clause and figure out where the degree variable belongs (in the subject), and only then do you stick it in in the corresponding spot in the second clause.

    If there's a dramatic mismatch between where the gap is and where the degree variable is, it might go unnoticed if the two are processed separately.

    Also, I want to credit Jack Hoeksema with finding the example and including it in his corpus.

  25. John Cowan said,

    December 27, 2009 @ 8:22 pm

    I think the sentence is perfectly cromulent in every way.

  26. Adrian Bailey said,

    December 27, 2009 @ 8:43 pm

    Took me a little while to get it. Could one call the "they" a dangling pronoun? One corrected version is "More people found Mr Bush's ads negative than found Mr Kerry's so."

    I enjoy it when such examples crop up: utterances that don't bear close analysis but are still in a sense grammatical ("natural" grammar, if you will) in that our brains have little or no problem interpreting them.

  27. Neal Goldfarb said,

    December 27, 2009 @ 11:54 pm

    A similar phenomenon is discussed in Gibson, E. & Thomas, J. (1999). Memory limitations and structural forgetting: The perception of complex ungrammatical sentences as grammatical. Language and Cognitive Processes, 14, 225-248.

    From the abstract:

    Results from an English acceptability-rating experiment are presented which demonstrate that people !nd doubly nested relative clause structures just as acceptable when only two verb phrases are included instead of the grammatically required three. Furthermore, the experiment shows that such sentences are acceptable only when the intermediate verb phrase is omitted.

  28. Fernando Pereira said,

    December 27, 2009 @ 11:55 pm

    It's been a long time since I worked on ellipsis, but the only way this example seems baffling is if "they" is expected to be referential. In the work on ellipsis that I did with Mary Dalrymple and Stuart Shieber, we took a more liberal view that explained fairly well several related puzzles. It this case, I'd take the matrix "they did _'s" as a syntactically well-formed pro-form for a to-be-determined higher-order function F that applies to "Kerry". Taking "more" to have type (e->t)->(e->t)->t, interpreting its first argument (simplified) as P = \x.(person(x)&found_negative(x, ads(Bush))), and setting the ellipsis-resolving equation as F(Bush) = P, the non-vacuous solution is F = \y.\x.person(x)&found_negative(x, ads(y))), yielding the overall interpretation more(\x.(person(x)&found_negative(x, ads(Bush))),
    \x.person(x)&found_negative(x, ads(Kerry))))

  29. Mark F. said,

    December 28, 2009 @ 12:01 am

    TL — The sentence you prefer, "In Michigan and Minnesota, more people found Mr Bush's ads negative than Mr Kerry's," probably is indeed the best way to phrase it. Some sentences of that form can be given an alternate interpretation: "More people like Madonna than Dan" might be willfully misinterpreted to mean "More people than just Dan like Madonna", and you can avoid that by repeating the "like". But even there there's no real ambiguity (I know it's not a very good example), and in the original sentence there's no ambiguity at all.

  30. Army1987 said,

    December 28, 2009 @ 8:18 am

    I can't make head or tail of the sentence as it's written, so I think that either:
    * the "they" is spurious, and Kyle's interpretation was intended, or, slightly less plausibly
    * there should be a missing "more" in the second clause, and it's intended to mean "there is a greater number of people in Michigan and Minnesota [than elsewhere] who found Mr Bush's ads more negative than they found Mr Kerry's ads negative. (Such a nested comparative is weird, but in a de-facto two-party system such as the American one virtually any judgement of a main candidate is a comparison with the other main candidate.)

  31. Franz Bebop said,

    December 28, 2009 @ 8:35 am

    I'm reminded of this little trick:

    http://help.com/post/220484-this-proves-spelling-doesnt

    Perhaps some syntactic processing works the same way. Just ignore the details and settle on something plausible.

    [(myl) This is meaningless advice, since it ignores the critical question of what makes something "plausible". In the case of spelling, the notion (advanced in the link you cite) that all other letters can be in random order, if only the first and last one are correct, is trivial to prove wrong, as discussed here. That same link will explain that the alleged "research at Cambridge University" is an urban legend.

    In the sort of case under discussion in this post, "ignore the details and settle on something plausible" not only fails to specify what is "plausible", it doesn't explain why some examples are accepted without a murmur, while others immediately cause readers or listeners to note a problem.

    This is like explaining the whole inventory of visual illusions by saying "your eyes and brain just ignore the details of the image and interpret it in a way that makes sense". That might be true at some level, but as a theory, it's somewhere in the territory between circularity and nonsense.

    You really need to be careful of believing twaddle that happens to coincide with your prejudices.]

  32. John Williamson said,

    December 28, 2009 @ 6:10 pm

    One solution is to use a noun which sets up a a comparison and then leave off most of the comparative apparatus:

    "In a Michigan and Minnesota comparison of ads of Bush and Kerry, more people found Mr Bush's ads to be negative.”

    In context, even this would work. "In Michigan and Minnesota, more people found Mr Bush's ads to be negative."

    From the word "more", the relatively lesser finding of negativity in Kerry's ads is clearly implied.

  33. Franz Bebop said,

    December 29, 2009 @ 4:10 am

    @Mark: Ha, yes. Thanks for setting me straight.

  34. vanya said,

    December 29, 2009 @ 11:24 am

    "Complete and utter syntactic nonsense."

    But of course it's not syntactic nonsense, if I and many others apparently grasped the meaning of the sentence immediately, and it took Andrew "days" to figure out what was wrong with it. That's just nitpicking to no purpose. In many dialects of English having a pronoun duplicate an antecedent subject is within the bounds of acceptable spoken discourse. It's Geoff's theories of syntax that have failed here.

    [(myl) You might be right; and then again, this might be the grammatical equivalent of an optical illusion, and you might be complaining that you and many others grasped immediately what the picture represents, and it took others special interventions (like making measurements or obscuring parts of the image) to see what's wrong with their immediate perceptions. How could you ever tell, if you assume that your immediate perceptions are always correct?]

  35. John Thayer Jensen said,

    December 29, 2009 @ 2:04 pm

    I have been racking my brain trying to remember who (it was a linguist) said something amounting to:

    "There is no such thing as syntax; there is only analogy."

    It might explain why people have no difficulty with things like "would of" instead of "would have."

  36. John Thayer Jensen said,

    December 29, 2009 @ 2:19 pm

    Vanya said:

    In many dialects of English having a pronoun duplicate an antecedent subject is within the bounds of acceptable spoken discourse.

    In that case I didn't understand it. I didn't take 'they' as referring to the same people as the 'more people' I thought it referred to a rather abstract 'people who thought X's ads negative' – the first a group of people who thought Bush's ads negative, the second – the antecedent of 'they' – a different group of people who thought Kerry's ads negative.

  37. Melanoman said,

    December 30, 2009 @ 5:15 am

    This sentence is fully grammatical and semantically valid in my native dialect. Let's isolate the most important elements.

    "In a recent survey, more people liked Pepsi than they did Coke."

    I take this sentence to be equivalent to:

    "The percentage of people in a recent survey who expressed a preference for Pepsi exceeded the percentage of people in that survey who expressed a preference for Coke."

    The less interesting (to me) change in the translation is from "like" to "prefers." I see that the sentence takes for granted an assumed model of how the survey works.

    The assumed model of the survey in this example is that of a taste test where the person indicates a preference between the two drinks. This is a little different than the original example where the assumed model is that the survey gives the person a chance to characterize each group of ads, and that one or more of the responses counts as negative. In both cases, if the hearer were to take a different model of the way the survey works, the derived meaning changes to match the model. If the models of the speaker and hearer do not match, a misunderstanding is likely.

    A more interesting change to me involves the verb "did." Did what? In this example, the predicate is "to express a preference for " and the phrase "…did Coke" uses a structure where the predicate is the same and "Coke" fills the valence . In the original example the predicate is "found to be negative." This works almost the same, except it adds the extra twist that "Mr. Kerry's" is a shortening of "Mr. Kerry's ads" in context.

    That leaves the frustrating "they." I we treat "they" like a pronoun with a rigid referent then "…they did Coke" would have "they" refer to the subset of people on the survey who liked Coke. The problem is that if we believe that the rigid reference drawn from the context, the only compatible choice in the context is "More people" which refers to the people who liked Pepsi — exactly the opposite of the people who liked Coke.

    My theory is that the same structure which allows the word "Coke" to replace the word "Pepsi" in the predicate (or "Mr. Bush's ads" with "Mr. Kerry's [ads]") also affects the interpretation of "they." The syntax of the predicate form requires a survey subset as a subject, and the semantic of the fully qualified predicate selects which subset.

  38. Melanoman said,

    December 30, 2009 @ 2:00 pm

    In future I need to use [x] for my valences instead of the angular brackets, which vanish as html tags. Some year I want to take a large block of time away from work and play linguist again. These snippets are just enough to whet the appetite but not enough to quench it.

  39. Fiestoforo said,

    December 30, 2009 @ 2:25 pm

    I think Melanoman is right, one could think that the pronoun "They" is syntactically referring to "Most People", when it is referring to "people". I would paraphrase it as:

    In Michigan and Minnesota, Mr. Bush’s ads were considered more negative than Mr. Kerry’s.

    rgs.

  40. John Thayer Jensen said,

    December 30, 2009 @ 3:13 pm

    My reading was exactly the same as Melanoman's and Fiestoforo's. I did not think the same people were under consideration in the two cases.

    If I understand Vanya correctly, he thought they were the same group of people, expressing an opinion about two different sets of ads.

  41. Aaron Toivo said,

    December 30, 2009 @ 5:10 pm

    Like most others, I judge it grammatical.

    But… exactly what does it even mean, when a comparison standard uses an anaphor for an antecedent located in the comparative clause? I think the best answer, i.e. the one that fits my instinct, is to follow an inferential approach and say that because they cannot logically have the same referent, we then fill in the only referent that does makes sense.

    Is there any good reason in cognitive linguistics why such a situation could not become a standard part of the grammar?

  42. Dan M. said,

    December 31, 2009 @ 1:40 am

    I feel the sentence is entirely grammatical and understand in almost exactly as described by Fernando Pereira, with the modification that the "they" being used explicitly to define P implies that the predicate is applied over the same body of people for both Bush ads and Kerry ads.

    That means that I find most of Mark Anderson's propositions well-defined:

    Either (1) or both (2) and (3) hold.

    (4) holds if by it Mark meant that the set of N people asked about Bush ads was the same set of N people asked about Kerry ads. That entails (5) and (6).

    (7) is exactly the main claim of the sentence.

    (8), (9), and (10) are NOT what the sentence says and are undecidable from the given information.

RSS feed for comments on this post