Keith Chen animated

« previous post | next post »

Jason Merchant sent me a link to this animation of Keith Chen's ideas about tense marking and future-orientation in financial and health behaviors:

Jason's note:

There are many errors in the reporting of Chen's paper in this animation (mostly overstating the conclusions and controls), but overall the animation is pretty impressive. I guess you know your idea has made it if someone makes an animated movie out of it featuring you. (Because Chen did not control for cultural factors though, it remains at best a supposition that language, and not the cultures of the people using them, are responsible for the savings and other behavioral differences found. In the two cases I know best of his 5 intra-country subcases, Belgium and Switzerland, it is simply a sociological error to believe that the Flemish and Walloons (bzw. German-speaking vs French-speaking Swiss) are culturally uniform and that they differ *only* [as eg the video claims, and as is implicit in Chen's paper] in their languages. I personally plan on doing a follow-up study showing that knowledge of drinking songs increases proportionally with savings rate (since German-speakers culturally have more drinking songs than French-speakers, including in Switzerland, this will show the same correlation that language does with savings rate: once again showing that correlation is not causation, the Achilles' heel of Chen's work. Hopefully this will lead to more people learning drinking songs?)

Thought you might enjoy this nonetheless–the meme is loose, and while the media did its transitory best to spread it, there's nothing like a well-done YouTube animation to influence people for all time…

I'm more persuaded by Chen's controls than Jason is,  though of course I agree about the difficulties of separating correlation and causation, and the need for more field research on drinking songs.

With respect to the video, I had a hard time getting past the pronunciation of Sapir as "supper", but I might have more to add after further viewing.

Previous LLOG coverage of Keith's work:

"Keith Chen, Whorfian economist", 2/9/2012
"Cultural diffusion and the Whorfian hypothesis", 2/12/2012
"Whorfian Economics", 2/21/2012
"Thought experiments on language and thought", 2/22/2012
"Keith Chen at TED", 2/20/2013

For those readers who might not be familiar with Edward Sapir and the pronunciation of his name, the Wikipedia article gives it as /səˈpɪər/, which rhymes with "a seer".


  1. Stan said,

    September 7, 2013 @ 7:35 am

    I notice the narrator says economic professor whereas the animated text has economics professor. Unlike similar pairs such as math/maths, economic is potentially ambiguous, albeit very unlikely to be misinterpreted in most contexts.

    Not to criticise, but I'm curious: is this use of economic professor standard AmE?

    [(myl) I don't recall hearing it before; nor "physic professor", "linguistic professor", etc.]

  2. Ellen K. said,

    September 7, 2013 @ 7:51 am

    I wonder, are their languages that really do always force a grammatical distinction between present and future? English, despite what the video says, does not. Spanish, which has a future tense, does not. Spanish even uses (at times) the simple present for future events (whereas English usually requires the progressive). Languages making a distinction between past and present isn't as either/or as this video suggests.

  3. Jerry Friedman said,

    September 7, 2013 @ 8:18 am

    Stan: Here are some COCA results:

    professor of economics: 369
    economics professor: 340
    economic professor: 3

    If I saw "economic professor", I'd think it was a typo.

  4. Bill Benzon said,

    September 7, 2013 @ 8:20 am

    "… I had a hard time getting past the pronunciation of Sapir as "supper"…"

    Maybe it's another one of those Eskimo words of snow.

  5. marie-lucie said,

    September 7, 2013 @ 8:35 am

    Edward Sapir and the pronunciation of his name, the Wikipedia article gives it as /səˈpɪər/, which rhymes with "a seer".

    This sounds British to me (like many Wikipedia transcriptions). In North America I have always heard the name rhyming with "appear", with the second half as "peer" (one syllable), not "seer" (two syllables).

    [(myl) In my idiolect, "seer" (at least in the sense I meant it) is one syllable, rhyming with "peer".

    And in r-ful varieties of English, the transition from /i/ to /r/ passes through a region of the vowel space that you could choose to transcribe as a a schwa-like offglide, as whoever crafted the Wikipedia pronunciation has done. I prefer just something simpler and more phonemic like [iɹ], not getting into the details of whether the nucleus has been affected enough by the /r/ to write it as [ɪ], or whether the transition to [ɹ] is salient enough to motivate interpolating some kind of rhoticized central vowel.

    At this point, though, we're into the territory where the topic become "why the IPA is an inadequate way to represent vowel quality"…]

  6. DaveK said,

    September 7, 2013 @ 8:45 am

    I've never heard "seer" pronounced with two syllables–it's always identical to the verb describing what you do to a steak.

    [(myl) Marie-Lucie can speak for herself, but I thought she meant the re-created agentive "see-er" (as in "the see-er and the seen"), which would be two syllables for me. But in general, English syllables with /l/ and /r/ in the rhyme (like "hour" or "dowel" or "seer" or etc.) are cases where the whole idea of syllable counting loses incoherence — there are some pronunciations that that seem to be clearly classifiable as monosyllabic or disyllabic, but there's also a broad region of uncertainty in between. And imaginary pronunciations (i.e. what it seems to you that you would say if you said it) are, as always, even less reliably classifiable than real ones.]

  7. Victor Mair said,

    September 7, 2013 @ 9:19 am

    The main pronunciation key here (link below) gives "seer" as one syllable, but the recording is ambiguous, though it really sounds to me as though it has two syllables, while the pronunciation key for the first definition specifies two syllables.

    I always say the word with two syllables, no matter what the meaning (except in the mathematical sense, where it is a variant of "ser", a Hindi unit of weight).

  8. Stan said,

    September 7, 2013 @ 9:29 am

    Mark, Jerry: Thanks. I figured it was non-standard based on preliminary searches (in corpora and on .edu sites via Google), but I didn't want to assume it wasn't an unusual but legitimate variant.

  9. xyzzyva said,

    September 7, 2013 @ 11:17 am

    That transcription highlights the diaphonemic nature of Wikipedia's English IPA. The idea is to keep the transcriptions dialectally-neutral when possible, so that a reader has the information necessary to construct a pronunciation in their own accent, but it often has the perverse effect of listing a pronunciation actually used by no one.

    And I dare you to try making a verifiable decision whether to use the cot or caught vowel for someone only ever spoken about by the cot-caught merged.

  10. Victor Mair said,

    September 7, 2013 @ 1:35 pm

    It's hard to believe that people are still taking Keith Chen's ideas seriously, even to the point of making a gushingly enthusiastic animated exposition of them.

    Although they are respectful of the formal features of Keith Chen's research, perhaps because it adheres to the trappings of mainstream economic theory and statistical methodology, and because he teaches at an Ivy League university, serious doubts about Chen's conclusions have been raised by my Language Log colleagues Geoff Pullum, Mark Liberman, and Julie Sedivy (references may be found at the end of the original post above).

    It would seem, moreover, that specialists have found fatal flaws even in the way Chen utilizes their analyses. For example, in this comment to one of the Language Log posts cited above (, Östen Dahl — upon whose data Chen draws heavily — sharply distances himself from the extrapolations of the latter.

    I've written a longish blog critiquing the purely linguistic aspects of Chen's work, specifically that a supposed lack of futurity in Chinese CAUSES Chinese people to have less unprotected sex, to smoke less, to gamble less, etc. — in comparison with English speakers who allegedly have more unprotected sex, smoke more, gamble more, etc. because their language has a future tense:

    "Futurity in Chinese and English and Its Supposed Economic Consequences"

    There is good follow-up discussion in the comments that merits careful reading.

    Aside from the inaccuracies with regard to linguistics pointed out in my Sinoglot post, my main quarrel with Chen's conclusions — no matter how impressive his statistical and economic apparatus — is that they simply do not comport with the empirical evidence.

    Let us take indulgence in eating as a test case. For those who can afford it, overeating is endemic in China (just go to a few Chinese banquets and you'll know what I mean), and obesity is a very serious problem, especially among young people. But I would like to focus on one particular disease that is related to an overly rich diet, namely, diabetes. Right now, China is facing an epidemic of diabetes:


    "China 'Catastrophe' Hits 114 Million as Diabetes Spreads"

    Bloomberg-Sept. 4, 2013
    China's rising prevalence of diabetes has strained its health services and helped fuel a 20 percent-a-year growth in drug sales, stoking the ….


    I am always amazed at how many people I meet in China who have diabetes, and the numbers continue to rise. Now, this must be the result of overindulgence and other bad eating habits, quite the opposite of what Chen predicts.

    Those who suffer from diabetes in the United States amount to 25.8 million people, 8.3% of the population. The percentage of those afflicted with diabetes in China is already higher than that figure and is increasing at an alarming rate. This has nothing to do with a change in the genetic pool or a shift in language, but is surely the result of increasing wealth and the bad habit of overeating rich food for the moment, future be damned.

    "China Diabetes Rate Now Higher Than U.S."

    "China is now home to the world’s largest diabetes population. The number of people who have diabetes or early signs of the disease is greater than the entire population of the U.S."


    Medical practitioners tell us that type 1 diabetes is genetically determined, but that type 2 diabetes is caused both by genetic and dietary factors. Medication is not advised for type 2 diabetes, which can usually be cured by changing away from overly rich diet.

    All right, enough about the consequences of Chinese eating too much rich food.

    They also smoke like chimneys:

    China male 59.5% female 3.7%

    USA male 26.3% female 21.5%

    Anyone who has spent any amount of time in China would surely realize that there are smokers everywhere one turns, both in the city and in the rural areas.

    Conversely, in America one sees fewer and fewer smokers. I very seldom encounter smokers on the Penn campus. The one person I know who smokes the most and is completely addicted to cigarettes is a male from Taiwan.

    Gambling? It's far worse in China than in America, and gambling may be the most dangerous addiction of all (I won't get into alcohol in this brief reply to Chen, but will only say that, except in rarefied circles, if you don't drink hard and heavy in China, you are considered weird).

    Macao has now displaced Las Vegas as the gambling capital of the world. Those who flock to the gigantic casinos in Macao are local and regional gamblers, but they come in huge droves from Mainland China as well.

    Singapore and Hong Kong rank among the world's biggest gambling centers. I would encourage Keith Chen to visit the casinos in those two places before writing about the effect of lack of futurity in Chinese and Finnish. By the way, 41% of adult Finns gamble every week. (see all the photos and read the whole of the accompanying texts to the right)

    Cf. also here:

    When Chinese visitors come to Philadelphia and you ask them where they want to go, more often than not they will say Atlantic City and Niagara Falls (!). Ditto for LA — Las Vegas and the Grand Canyon (!) — and off they go tearing across the desert.

    Gambling is particularly prevalent among Asian Americans, especially those of Chinese ancestry, precisely because of cultural reasons.

    Literature stretching back for millennia also attests to the Chinese love of gambling. My favorite recent author, Lu Xun, has some great stories about gambling fanatics, but you can find such tales in all eras of Chinese history.

    My own Chinese father-in-law was addicted to gambling, which almost ruined his family, and the only thing that saved him (and the family) was conversion to Christianity (he stopped gambling immediately and totally once he became a fervently practicing Christian).

    The Chinese propensity for gambling is not something that came into being after Westernization, but it has very deep roots in history. Google on: gambling in chinese culture // gambling in chinese history

    Sex? I could say an awful lot about license in that regard (how about back rooms for sexual services in a very large proportion of barber shops and calls to your hotel room in the middle of the night?), but I will mention only that a friend of mine actually heads the operations of a major condom provider in China and his biggest challenge is getting people to understand the dangers of engaging in unprotected sex. Thus, as much of his efforts go into education as into marketing. This obviously has enormous consequences for the transmission of HIV-AIDS, and I also know a lot about that too because another friend of mine heads an NGO dedicated to diminishing the worrisome levels of this deadly disease.

    In short, the results of Keith Chen's research are divorced from reality; his findings evince a terrible disconnect between theory and actuality

    I submit that, while Chen's methodology may be elegant, his conclusions are bunkum and not at all a good test of the Sapir-Whorf hypothesis. I was going to call his proposals "voodoo economics" or "zombie economics", but I found that those expressions have already been used to designate different kinds of economic fallacies, so I refer to Chen's variety as "mumbo-jumbo economics" instead.

    I'm not sure I'd ever accept an invitation to appear on TED, since so much of what is presented there is sheer schlock. Come to think of it, I don't think I'd want one of these animated lectures done for my work on mummies, Chinglish, character amnesia, or biànwén 變文 ("transformation texts") either, because I'd be afraid that the presentation would be so hyped and distorted that I myself would barely recognize what it was about.

    As for Sapir becoming "supper", from the very first sentences I had a strong suspicion that the voice was machine generated.

  11. D Sky Onosson said,

    September 7, 2013 @ 5:06 pm

    I couldn't make it past the claim that "It is cold tomorrow" is not possible in English.

  12. J Lee said,

    September 7, 2013 @ 5:30 pm

    myl: '…are cases where the whole idea of syllable counting loses incoherence '

    you should do a post about double negatives

  13. chris said,

    September 9, 2013 @ 6:27 am

    Marie-Lucie can speak for herself, but I thought she meant the re-created agentive "see-er" (as in "the see-er and the seen"), which would be two syllables for me.

    But that's exactly what a "seer" is: one who sees, just like a watcher watches, a fighter fights, and a driver drives. It wouldn't have occurred to me that some people would pronounce it the same as "sear". Although the general point about syllable counting is a good one: ambiguity is possible and attempts to declare it out of existence, so to speak, tend to result in arbitrary line-drawing.

    It's hard to believe that people are still taking Keith Chen's ideas seriously, even to the point of making a gushingly enthusiastic animated exposition of them.

    In my experience, there are few, if any, ideas so foolish or unsubstantiated that *someone* won't take them seriously. Examples from politics and religion are legion, but it would perhaps be rude to name specifics.

  14. Victor Mair said,

    September 9, 2013 @ 8:34 am


    "In my experience, there are few, if any, ideas so foolish or unsubstantiated that *someone* won't take them seriously. Examples from politics and religion are legion, but it would perhaps be rude to name specifics."

    The problem with Keith Chen's whacky ideas is that it's not just *someone* taking them seriously, but LOTS OF PEOPLE doing so.

  15. Victor Mair said,

    September 9, 2013 @ 10:18 am

    Me to David Sapir, emeritus professor in the Department of Anthropology at the University of Virginia:


    There's a discussion going on in the comments section here about how your father pronounced his surname:

    "Keith Chen animated"

    I, and also I think most people, pronounce it as "saw-PEER".

    Presumably you pronounce your surname the same way your father did. Please enlighten us.


    Prof. Sapir to me:


    sa sounds as su as in supper

    pir sound like pier as where you tie up a boat.

    accent on the second syllable.


  16. J.W. Brewer said,

    September 9, 2013 @ 12:07 pm

    I think I've been slightly mispronouncing "Sapir" for quite a long time and am indebted to Prof. Mair for taking the initiative to get evidence from a primary source. Now of course I'm worried that I've been mispronouncing "Whorf" for an equally long time. I treat it as a homophone of "wharf" (which overlaps semantically with "pier," as in the pronunciation of "Sapir" -certainly no mere coincidence?). I have little if any w/wh distinction in my speech, so it's really a question of whether I've got the right vowel. Any contradictory views?

  17. Victor Mair said,

    September 9, 2013 @ 12:24 pm

    Well, now, if anyone knows a descendent of Benjamin Lee Whorf, please ask them how they say their surname. I pronounce it with a breathy "w" and the vowel sounding like the "o" in "more".

  18. Jonathon Owen said,

    September 9, 2013 @ 12:44 pm

    "the Flemish and Walloons (bzw. German-speaking vs French-speaking Swiss)"

    Shouldn't this be Dutch-speaking vs. French-speaking Belgians? I'm sure this was just a braino, but I'm surprised nobody has pointed it out.

  19. J.W. Brewer said,

    September 9, 2013 @ 1:30 pm

    My reading in context is that Jason was describing two different situations (one Belgian, one Swiss), although it may have been ill-advised to use a German abbreviation ("bzw") which I myself could not successfully summon up from my rusty high-school German and which wikitionary suggests may be confusingly polysemous.

  20. J.W. Brewer said,

    September 9, 2013 @ 1:34 pm

    I should probably be respectful and say Prof. Merchant instead of "Jason," especially after a review of his CV shows just how much better he did in his undergraduate linguistics studies a few years after me at our common alma mater (from which I infer that he probably did not spend nearly as many hours at the campus radio station as I did).

  21. Jonathon Owen said,

    September 9, 2013 @ 2:05 pm

    J.W. Brewer: Yes, I think you're right. I suppose the braino was my own.

  22. J Lee said,

    September 10, 2013 @ 3:13 am

    "It wouldn't have occurred to me that some people would pronounce it the same as "sear". "

    like marie's peer/seer pair (supposedly 1 & 2 syllables) i dont see any basis for this comment. isnt this an empirical question subject to certain principles, e.g. there is no motivation to transcribe off-glides since it is merely a phonetic detail due to the distinct articulation of the English rhotic? would anyone say 'searing' has 3 syllables, although it contrasts with the phrase 'see a ring' where a glide is necessary and audible?

    nor do i understand the other possible pronunciations of foreign names with the form CvCiC. how an english speaker could settle on 'supper' for 'Sapir' is baffling. One *'sapir' /seipr/ could be modeled on tapir (Fr. /tapir/) and it seems to me that reducing the first vowel of these names is optional.

    i would be interested to know if this pronunciation /Cə'CiC/ became widespread because of the fake muslim names black people have inflicted on their children since the 60s

  23. Victor Mair said,

    September 10, 2013 @ 5:52 am

    This is an addendum to my long note above (

    Never mind that Keith Chen's proposals do not jibe with empirical evidence (smoking, gambling, excessive ingestion of rich foods, etc.), they are counter-intuitive from their very premises. It would make much more sense for someone whose language has a strong sense of futurity (i.e., built into the grammar) to be more concerned about the future consequences of bad diet, unprotected sex, smoking, and so forth, than someone whose language lacks a strong sense of the future (think and live for the present). But the whole argument is specious anyway — grammar does not determine behavior — and this is not what Sapir-Whorf were talking about.

  24. Jean-Pierre Metereau said,

    September 10, 2013 @ 1:49 pm

    I've always been intrigued by the Sapir-Whorf hypothesis, and I was excited to see it addressed. From the initial mis-pronunciation of Sapir to the blatantly mistaken assertion pointed out by D Sky Onosson, however, I realized that this animation wasn't going to do much good for the reputation of the SW hypothesis. English does not really have a future tense the way French, for example, does. But it has several ways of indicating the future. And so does Finnish. The idea that a particle (such as the French -ra) is enough to make Francophones more liable to indulge in risky behavior than Anglophones is ridiculous. This is not the way to rehabilitate the much-battered SW hypothesis.

  25. M.N. said,

    September 10, 2013 @ 7:17 pm

    (When I was in high school and saw "Sapir-Whorf" written down for the first time, I actually did think that "Sapir" rhymed with "tapir"!)

    To add to Jean-Pierre's point above, doesn't the very existence of expressions like "tomorrow" undermine the hypothesis that lack of future marking makes the future somehow seem more like the present to speakers of Finnish-type languages?

    Also, I thought I was the only one who used "bzw" in English! :-) When I was in Germany recently, I saw it translated as "resp." in multilingual signage, but that's not an abbreviation I've otherwise seen used in English. Maybe they use it in Britain? I've never seen it in the US, or in books. But "bzw" is super useful, and I don't know how I survived without it. :-)

  26. Derek said,

    September 10, 2013 @ 9:46 pm

    @Victor Mair: This was my first thought too. When the narrator was giving the introduction about futured and futureless languages, I was expecting the thesis to be that futured speakers better prepare for the future, while futureless speakers live in the present. I was surprised to see the opposite claim.

    Regarding "Sapir", I always pronounced it with a short 'a'. I'm glad to know the correct pronunciation now.

  27. Victor Mair said,

    September 11, 2013 @ 7:18 am

    If lack of grammatical futurity were such a panacea for much of what might potentially ail people, China and Finland would not have among the highest suicide rates in the world. People who commit suicide tend to have bleak thoughts about the future.

    Except perhaps for Bangladesh, China appears to be the only country in the world where the rate of suicide among women exceeds that among men.

  28. Ellen K. said,

    September 11, 2013 @ 12:44 pm

    I get the logic of the claim. The idea that grammatical ways to mark the future (not necessarily a future tense, at least in this video's wording of the claim) separates out the future and makes it seem more distant. I also get the logic of the opposite viewpoint. Which is why logic isn't enough, you have to look at data. Which supposedly this study did. But did it do so accurately?

    The whole idea of separating languages into either languages that have a grammatical future and those who don't seems too simplistic to me, based on my familiarity with the two languages I speak, Spanish and English. It's not as simple has languages either having a future tense that you always use to mark the future, or else not marking the future, as this study seems to say.

  29. Victor Mair said,

    September 13, 2013 @ 6:01 am

    Hard liquor more important than human food or animal food.

    "China Turns to US Sorghum for Animal Feed"

    China's large private feed mills are increasingly buying sorghum from the US after using up their import allocations for corn, the preferred animal-feed grain.

    In China, sorghum is traditionally used to produce liquor, such as the potent baijiu rice wine. Now China is using the grain to help formulate animal feed. China's large private feed mills are increasingly buying sorghum from the US after using up their import allocations for corn

RSS feed for comments on this post