Sharks and New Yorkers

« previous post | next post »

Andrew Dowd, the intrepid and renowned hunter of syntactic rarities, has finally managed to find a "Russia sentence" captured on video. There are bound to be some of you out there who didn't fully believe that people walk around saying nonsensical things like More people have been to Russia than I have. You needed harder proof. Well, here it is, on YouTube, and what's more, the speaker is an experienced broadcaster confidently compering a TV panel game:

At around 30 seconds in, you can hear and watch Stephen Fry (he of the unbearably prissy recent BBC language program Fry's Planet Word) say: It so happens that more people in the world are bitten by New Yorkers every year than they are by sharks.

Think that through with your grammar module fully powered up and you'll see that it is not a sentence and couldn't possibly mean anything. Yet Fry thinks it does, and the panel members seem to understand it, and so does the whole studio audience. It's so weird: a total non-sentence (perhaps we should call them "nonsences"?) that everyone is tempted to think, wrongly, is just fine: as I put it a couple of years ago, when we look back at such a sentence we realize "it couldn't possibly be claimed to have the right syntax to say what we (wrongly) thought it said."

Linguists do not understand this phenomenon. But they are keeping it under observation.

Material added 2 May 2012, 7 a.m. EST:

Work on the topic is being done at the University of Maryland, by Colin Phillips and other linguists and psycholinguists, among them E. Lau, Matt Wagers (now at the University of California, Santa Cruz), and Alexis Wellwood. See, for example, C. Phillips, M. Wagers & E. Lau (2009), "Grammatical illusions and selective falibility in real time language comprehension", in Jeff Runner (ed.), Experiments at the Interfaces, volume 37 of Syntax & Semantics, Emerald Publications. Phillips et al. have run a number of experiments to try to get to the bottom of how the illusion arises.

Colin Phillips points out to me (by email) that the Stephen Fry example (More people in the world are bitten by New Yorkers every year than they are by sharks) is not quite the same sort of thing as the stereotypical Russia sentence (More people have been to Russia than I have), and I agree: I noticed the difference, but felt the two had enough in common to count of instances of a single phenomenon. It's been suggested that the term linguistic illusion might be useful in talking about the relevant class of phenomena in broad terms, and that is how the Maryland researchers are referring to it. Phillips goes on:

The problem with the genuine Russia sentences is that plural subject in the first clause ("more people") has no corresponding countable NP in the second clause to be compared to. On the surface, Stephen Fry's example is similar ("More people … they"). But since "they" is anaphoric to the bare plural "people", it does not fail in the same way that the classic Russia sentences do. What is interesting about the Fry example, I think, is the way that that pronoun is anaphoric to something that it is not intended to corefer with.

For what it's worth, our experiments on Russia sentences have tested a number of popular ideas about the source of the illusions, e.g., mis-recall of word order due to category ambiguity of "more", the "not just me" interpretation, etc. Most of them turn out to find little support in the experiments. The one thing that does seem to consistently matter is whether the predicate is repeatable/countable, e.g., an individual can go to Russia multiple times. Non-repeatable predicates consistently reduce the illusion effect. So the illusion seems to be related to a phenomenon that has interested semanticists, where statements about quantities of individuals are understood as assertions about quantities of individuals, e.g., "100,000 cars drove across this bridge last month", which is not falsified if it turns out that some individual cars contributed many bridge-crossing events. We have a detailed poster on this work, and a paper on it is almost complete.


One other thing. Some people have written to me to insist that there's nothing wrong with the sentence Fry uttered, since "its meaning was clear". Yes, everyone understood him; that is the whole source of the puzzlement here. People who insist that the sentence is fine are (if I'm correct) in the position of people who insist that the first line really is shorter than the second in the Müller-Lyer illusion. Yes, we know you are sure of that. That is the phenomenon of interest. If you really think that More people in the world are bitten by New Yorkers every year than they are by sharks can mean "New Yorkers bite more people every year than sharks do", then I have to say that I think you're mistaken, and to begin seeing that you should ask yourself what they refers to.


[More people are prepared to open comments than I am. But they are open.]



  1. Keith M Ellis said,

    May 1, 2012 @ 4:10 am

    Just replacing "they" with "there" pretty much fixes it, doesn't it? Where "are people" following "there" is implied? Which is one possibility — that it's a word substitution in a complicated sentence.

    However, I think it's more likely that somehow "people in the world" is conceptually doing double duty; there's an equivocation of people bitten by sharks and people bitten by New Yorkers that doesn't literally make sense but maybe does at some cognitive level. Just as in the Russian sentence.

    Specifically, a set and its subsets aren't being explicitly disambiguated, but nevertheless are cognitively distinct and equivocating them as one abstract thing is a sort of shorthand that works because it's, as you say, not possible to parse the sentence at all without inferring what's missing; therefore the inference is forced?

  2. HKlang said,

    May 1, 2012 @ 5:57 am

    I'm pretty sure it's a hypercorrection, and that's what makes it so galling.

    But consider: more humans are bitten by sharks than Martians by quarks.

  3. steve, from the internet said,

    May 1, 2012 @ 6:14 am

    Actually, removing "they" fixes the sentence perfectly. More people in the world are bitten by New Yorkers every year than are by sharks.

  4. Andrew Jorgensen said,

    May 1, 2012 @ 6:43 am

    Is it the same kind of sentence? You can't fix the original Russia sentence by deleting the 'I have' the way you can fix the shark sentence by deleting the 'they are'. The original sentence purports to be a comparison but can't really be one. Maybe it means 'some people have been to Russia but I haven't'. The shark sentence means to compare the incidence of being bitten by a New Yorker to the incidence of being bitten by a shark.

  5. Marc said,

    May 1, 2012 @ 6:55 am

    What about the old on-air slogan from ABC News: "ABC News. Where more Americans get their news than from any other source." Does this count as a Russia sentence?

    It made an impression on me when I first heard it because it that had clearly gone through lots of review, focus groups, editing and approval. I remember a later version of the phrase that went like this: "More Americans get their news from ABC News than any other source". Maybe they caved to the peever lobby?

    A quick search on You Tube didn't reveal any primary video evidence for either phrase. But, Google showed several secondary sources in print for both.

  6. Faldone said,

    May 1, 2012 @ 6:56 am

    Keith M Ellis may be on to something in this example. Fry is non-rhotic, isn't he?

  7. richard howland-bolton said,

    May 1, 2012 @ 7:01 am

    Makes perfect sense to me: the people who are bitten by New Yorkers are extremely unlikely to then be bitten by sharks in the same year.

  8. Eric P Smith said,

    May 1, 2012 @ 7:01 am

    I think that Keith Ellis's suggestion of replacing 'they' with 'there' goes some way towards fixing it, by bleaching the pronoun of its unwanted semantic content. But as steve from the internet says, to fix the syntax, the pronoun has to be removed altogether. Fry intends, "More people in the world are bitten by New Yorkers every year than are by sharks," or "More people in the world are bitten by New Yorkers every year than by sharks."

    At least it is clear what meaning Fry intended. By contrast I think the intended meaning of "More people have been to Russia than I have" is far from clear. The speaker may intend "More people have been to Russia than I" (or in other words "Other people besides myself have been to Russia"), or alternatively the speaker may intend "People have been to Russia more than I have."

    To my ears, the more natural of the two interpretations of "More people have been to Russia than I have" is "More people have been to Russia than I." The speaker knows that "She has more charm than I have" is equivalent to "She has more charm than I," and adds a spurious "have" to "More people have been to Russia than I" by false analogy. And I suspect that Fry has been taken in by a similar false analogy. Fry knows that "Tom and Billy are more frightened by New Yorkers than they are by sharks" is equivalent to "Tom and Billy are more frightened by New Yorkers than by sharks". He has taken "More people in the world are bitten by New Yorkers every year than by sharks" and he has inserted a spurious "they are" by false analogy.

  9. Faldone said,

    May 1, 2012 @ 7:01 am

    I have now listened to the clip and it could well be that he is saying "… there are by sharks." BTW, on first hearing I didn't understand one word of the first bloke's answer.

  10. Giacomo Ponzetto said,

    May 1, 2012 @ 7:18 am

    Like Steve above I would have said More people are bitten by New Yorkers than are by sharks, though with the reduced confidence of a non-native speaker.

    But now I have novel doubts concerning the sentence: More people are bitten by New Yorkers than those who are by sharks.Is this sentence syntactically correct?Does it mean People bitten by New Yorkers are more numerous than people bitten by sharks?Or does it mean People bitten by sharks are not the only ones bitten by New Yorkers?Or is it grammatically ambiguous between these two meanings?

  11. Steve F said,

    May 1, 2012 @ 8:49 am

    @ Faldone:
    As someone who speaks English with roughly the same non-rhotic accent as Stephen Fry, I am fairly sure that he says 'they' not 'there' – even a non-rhotic accent would have a linking r in 'there are'.

    As for the 'first bloke' (Johnny Vegas), his Lancashire accent is entirely comprehensible to me, (and presumably most Brits, or else he wouldn't have a career) but I can appreciate why he might be difficult to understand: the reference to 'pork scratchings' (a delicacy that may not have world wide distribution) can't help.

  12. Mr Punch said,

    May 1, 2012 @ 10:14 am

    Like HKlang, I think this is hypercorrection. Being a prissy language maven puts one under considerable pressure.

  13. Eric P Smith said,

    May 1, 2012 @ 11:16 am

    Re Fry's rhotic accent, I'm with Steve F. Fry undoubtedly says "they" not "there". Firstly as Steve F says, Fry's accent would have a linking r in "there are" even though it is generally non-rhotic. Secondly, Fry's vowel in "they" is the diphthong /ei/, which is quite different from a British "there", rhotic or otherwise.

  14. Keith said,

    May 1, 2012 @ 1:49 pm

    I've tried turning my grammar module up to 11, and I think it is perfectly grammatical if understood as "each year, there are more people bitten by New Yorkers than there are New Yorkers bitten by sharks".

    That is to say, that in Fry's phrase, the word they refers to New Yorkers.

    Do you think that maybe my grammar module has a range of 1 to 20, and I need to turn it up higher?


  15. Joe said,

    May 1, 2012 @ 2:51 pm

    I was searching around in COCA, and found the following:

    "A lot more people have kids than they have capital gains" (Houston Chronicle 1999).

    It really is bizarre that we readily understand such sentences.

  16. Mark F. said,

    May 1, 2012 @ 3:04 pm

    When the US Transcontinental Railroad was being finished, the two teams working from opposite directions missed each other and overshot by some distance. Maybe this kind of sentence is the same sort of thing. One brain module was starting from the left, working on the sentence "More people are bitten by New Yorkers than by sharks" (leaving the second 'are' in there makes it sound awfully awkward to me). Another module was starting from the right, trying to construct the sentence "People are more likely to be bitten by New Yorkers than they are by sharks". When they met in the middle, they just shipped the sentence as was.

  17. D.O. said,

    May 1, 2012 @ 3:18 pm

    Maybe it's another instance of "poor monkey brains". More people is semantically quite a complicated construction. There is one group of people and then there is another one, with overlapping or disjoint membership, and the number of elements in the first group is larger than the number of elements in the second. That's hard. So when a sentence is simple enough, like "More people are ugly than stupid", the speaker is able to keep in mind that the sentence is about the relative number of people. But when it becomes more complicated, with a story line discussing visiting Russia or being bitten by sharks, it just slips speaker's mind that he talks about numbers and he starts talking about people themselves.

  18. HKlang said,

    May 1, 2012 @ 3:33 pm

    More New Yorkers eat sushi than Japanese drive Ferraris.

    *More New Yorkers eat sushi than New Yorkers drive Ferraris.

    *More New Yorkers eat sushi than they drive Ferraris.

    More New Yorkers eat sushi than drive Ferraris.

    The middle two sentences follow a paradigm that is perfectly logical, but the repetition is deprecated. Then again maybe I haven't got the stars right.

    More New Yorkers eat sushi than mudbugs.

    More New Yorkers eat sushi than Canadians.

  19. Carl Burke said,

    May 1, 2012 @ 4:52 pm

    I would agree that there are more New Yorkers who eat sushi than there are New Yorkers who eat Canadians, at least who will admit to it publicly.

    In the original sentence I would expect 'they' is just there to show agreement with 'more people', both plural and animate. Replacing 'they' with 'there' doesn't cover as much ground.

  20. Adrian Morgan said,

    May 1, 2012 @ 9:37 pm

    I'm with the "there's nothing wrong with this sentence" crowd (e.g. Keith). Even if you think it's clumsily-formed, I certainly don't see anything mysterious about its comprehensibility.

    Taking a stripped-down version of the original sentence gives: "more people are bitten by New Yorkers than they are by sharks". Inserting antecedents etc gives: "more people are bitten by New Yorkers than [people] are [bitten] by sharks". Converting to semi-mathematical syntax gives: "[people who are bitten by New Yorkers] > [people who are bitten by sharks]".

    In the original Russia sentence ("more people have been to Russia than I have"), conversion to semi-mathematical syntax gives: "[people who have been to Russia] > [I who have been to Russia]", which fails as a sentence primarily because "I who have been to Russia" doesn't correspond to a number in any sensible way. Replacing it with "more Americans have been to Russia than Australians have" gives "[Americans who have been to Russia] > [Australians who have been to Russia]", which seems fine to me.

  21. Chad Nilep said,

    May 2, 2012 @ 2:15 am

    The existence of "Russia sentences" suggests one of two explanations.

    1. Linguists' understanding of how human language works is just wrong, wrong, wrong. There is no internally-consistent, fully rule-governed grammar. People simply extract meaning from utterances by some combination of pragmatic understanding and probabilistic comparison with past utterances–more like Google algorithms than Cartesian homunculi.

    The obvious counter argument to 1 is that most parse-able utterances are meaningful, and most meaningful utterances are parse-able.

    2. Our minds, monkey-derived though they are, are incredibly flexible at deriving meaning from utterances, including poorly constructed ones. Something like coercion effects (to say nothing of pragmatics, probabilities, et alia) pushes us right past the impossibility of the unparseable to construct some meaning in spite of ourselves.

    Actually, though, I think I'll find, as Ben Goldacre might remind me, it's a bit more complicated than that. Although I've disposed of 1 as obviously countered, I do think there is something to say for theories of emergent syntax and the like.

  22. Michael said,

    May 2, 2012 @ 2:27 am

    "ןt's so weird: a total non-sentence (perhaps we should call them "nonsences"?) that everyone is tempted to think, wrongly, is just fine"
    — No mystery here; the listeners make the same mistake as the source. Analogous to the "No brain tumor is too small…" type verbal illusion.

  23. RI said,

    May 2, 2012 @ 5:34 am

    "Stephen Fry (he of the unbearably prissy recent BBC language program Fry's Planet Word)"
    –> Thank you. The program was infuriating. Obfuscation and misrepresentation in service of His Fryness! I love Stephen Fry but as a reasonably well-read amateur linguist this program really set my teeth on edge.

  24. Colin Phillips said,

    May 2, 2012 @ 9:02 pm

    @Chad Nilep: It's important to distinguish cases where linguists or prescriptive grammarians simply got the facts wrong from cases where speakers own initial judgments diverge from their more considered judgments. The interest of examples like the "more people have been to Russia …" sentences is not that people judge them to be ok, but rather that most people initially accept them, and then after further thought agree that they are nonsense (our group, led by Alexis Wellwood, has tested very many people on sentences like this). Visual illusions are cases where our perception mismatches facts about the external world. Linguistic illusions are cases where our own judgments/interpretations are inconsistent in a systematic fashion.

  25. Jason Merchant said,

    May 2, 2012 @ 10:11 pm

    Fry's sentence (and Joe's, above) seems to involve a simple resumptive pronoun ("they") were we would normally expect a gap–and this resumptive is the highest subject, and so patterns with Swedish (and Vata) resumptive pronouns, not Hebrew or Irish ones. (If work on Swedish is right, then this is a kind of "repair" strategy, where the violation being repaired is something like a comp-trace effect; wrongly, here, as "than" doesn't trigger this.)

  26. Chad Nilep said,

    May 3, 2012 @ 2:56 am

    @Colin Phillips

    I take your point, but I partially disagree. I think it is important *not* to distinguish too completely between linguists' models and speakers' "considered judgements". Each is a metalinguistic theory — fully powering up one's grammar module is a socially and cognitively specific activity, albeit a more ad hoc one than writing a descriptive grammar. As such, each is subject to error, or at least variation.

    That is not to say that linguistic illusions are not linguistically and psycho-socially interesting; they definitely are, as I'm sure you know better than I do. But neither side of that judgement mismatch, the rapid or the considered, the pragmatically coerced or the (meta-)grammatically informed, is clearly superior.

  27. Joe said,

    May 3, 2012 @ 2:29 pm

    If anyone is collecting these types of sentences, I found one in the BNC that also seems to be of the same type:

    "It is extraordinary that many more people have an inspection when buying a second-hand car than they do when purchasing a house"

  28. Abbie said,

    May 3, 2012 @ 6:30 pm

    Can I just defend Stephen Fry in general? QI is a fantastic show- nothing like it in America. This is a weirdly hostile attack on what is- at the very worst- a minor slip of the tongue.

  29. boynamedsue said,

    May 5, 2012 @ 5:02 am

    I agree with those who say that they/there is doing something here, though I'm not sure what. I see that Fry's sentence is not complying with any of the rules of English grammar. But…

    'More people are fat than they are lazy.'

    …sounds right to me too, and suffers from the same problem. Are we actually dealing with a set phrasal structure that exists outside of the normal rules of English?

  30. goofy said,

    May 7, 2012 @ 9:31 am

    Why is this sentence ungrammatical? What exactly, in syntactic terms, is the problem with it?

  31. Swimmy said,

    May 16, 2012 @ 3:15 pm

    @goofy, this comment is late enough that you probably won't see it, but the best way to understand is to replace the pronoun with its referent. Here, "they" is probably intended to refer to "more people" and not "New Yorkers." So the sentence reads, "More people in the world are bitten by New Yorkers every year than more people are by sharks. Since it's the same group of people, the comparison makes no sense. It says that there are more X than X. Hence humorous attempts to sort out the grammar by changing the intended meaning, a la richard howland-bolton and Keith.

    This is why replacing "they" with "there" fixes the sentence. "More people in the world are bitten by New Yorkers every year than there are by sharks" has a nonstandard structure, but is grammatical. It can be reformed into a standard structure by moving some words around: "There are more people in the world bitten by New Yorkers every year than by sharks." Here the comparison makes sense, because the sentence says there are more X than Y.

    @Abbie: I am not sure where you found hostility in this post. Fry made a grammatical flub, as we all do from time to time, and nobody has said that he's stupid, ignorant, or silly for doing so. We're focusing on it not out of anger or hostility but out of fascination that his ungrammatical sentence makes sense to almost everyone. The vast majority of possible ungrammatical sentences would sound like gibberish to us, so an ungrammatical sentence that we somehow parse well is rare and worth studying.

  32. Daz_Voz said,

    June 11, 2012 @ 7:19 am

    Fry is an experienced and elegant writer and I doubt that he would make a mistake like this in print.

    It was a slip. The sentence becomes conventional if you leave out the "they are", or (arguably) even if you leave out the last "they". Although an astute listener would detect the error, there really is only one thing he can have meant.

    This contrasts with the classic Russia sentence, which really can't be interpreted to mean anything. "More people have been to Russia than I have." What does the speaker actually intend?

  33. Ali stanton said,

    June 19, 2012 @ 11:13 pm

    its interesting stuff, but it seems to me to be quite similar to other illusions, in that you can only hold one meaning in your head at one time. when you look at a visual illusion it cant work as both a vase and two faces at the same instant. the mind has to perform a little flip for these two things to be seen.
    the same seems true of these sentences. for me, i was initially completely confused by the russia sentence, and couldnt get the intended meaning from it at all…then suddenly it happened and I have to stare at it hard again to get the nonesense meaning back.
    bytheway, It really helped to imagine the russia sentence being said in a really emotional tone of voice, for the sentence to turn from nonesense into a perfectly reasonable thing to say!

  34. Chad said,

    March 6, 2014 @ 5:50 pm

    I think the only problem is the second "they are". Take that out, and the sentence makes sense, albeit in a way that is still very silly.

  35. Davie said,

    June 13, 2014 @ 11:45 am

    @Faldone, that first guy is johnny Vegas and I understand him perfectly, maybe that is because I come from the UK so I am used to regional accents.

RSS feed for comments on this post