Julie Sedivy is the lead author of Sold on Language: How Advertisers Talk to You And What This Says About You. She contributes regularly to Psychology Today and Language Log. She is an adjunct professor at the University of Calgary, and can be found at juliesedivy.com and on Twitter/soldonlanguage. Keith Chen, an economist from Yale, makes a startling claim in an unpublished working paper: people’s fiscal responsibility and healthy lifestyle choices depend in part on the grammar of their language.
Here’s the idea: Languages differ in the devices they offer to speakers who want to talk about the future. For some, like Spanish and Greek, you have to tack on a verb ending that explicitly marks future time—so, in Spanish, you would say escribo for the present tense (I write or I’m writing) and escribiré for the future tense (I will write). But other languages like Mandarin don’t require their verbs to be escorted by grammatical markers that convey future time—time is usually obvious from something else in the context. In Mandarin, you would say the equivalent of I write tomorrow, using the same verb form for both present and future. Chen’s finding is that if you divide up a large number of the world’s languages into those that require a grammatical marker for future time and those that don’t, you see an interesting correlation: speakers of languages that force grammatical marking of the future have amassed a smaller retirement nest egg, smoke more, exercise less, and are more likely to be obese. Why would this be? The claim is that a sharp grammatical division between the present and future encourages people to conceive of the future as somehow dramatically different from the present, making it easier to put off behaviors that benefit your future self rather than your present self. Chen’s paper has yet to be accepted for publication, but it’s already generated a lot of press of the sort that’s festooned with flashing lights. For example, in his popular blog, Andrew Sullivan headlined the story with the pronouncement Why Greeks Haven’t Saved for a Rainy Day. A facetious headline, no doubt. But before someone suggests that the European Union should make bailouts of troubled countries contingent on their retiring their grammatical tense markers, it’s worth taking a reality check about the ways in which language can or can’t affect the thoughts and behaviors of its speakers. Claims about the tight coupling of language and culture are incredibly seductive. To many people, it’s intuitively obvious that dropping consonants in pronunciation is the mark of a lazy culture, that romancing someone is easiest in a language that’s intrinsically as soothing and soft as French, and that the disciplined German mind is in part a product of the strictly rigid and orderly German language. The trouble is, such intuitively obvious observations are bubbles just waiting to be burst by the sharp edges of actual linguistic evidence. As noted by Guy Deutscher, in his book Through the Language Glass, “the industrious Protestant Danes have dropped more consonants onto their icy, windswept soil than any indolent tropical tribe. And if Germans do have systematic minds, this is just as likely to be because their exceedingly erratic mother tongue has exhausted their brains’ capacity to cope with any further irregularity.” One of the most unkillable misconceptions is that if a language has no word for a particular concept, then its speakers must have trouble conceiving of it. Is Italian culture vulnerable to corruption because there is no Italian word that directly translates as accountability? Not likely. English doesn’t have a word for a silk green-and-pink paisley shirt, left untucked on one side, but I doubt that this makes it hard for you to picture one. Yet, this kind of thinking proves irresistible to many people—it’s even been used to argue that people who speak languages without future tense marking are unable to think about the future in any meaningful way. If this feels intuitively plausible to you, consider the following: In English, we mark gender on the third person pronouns he and she. But we don’t mark gender when we use pronouns to refer to a group of men or women—we use they in both cases. Does this mean that we suffer confusion about people’s gender as soon as they congregate in groups? Obviously not. And do languages that obsessively classify all nouns as gendered (as does Spanish) result in cultures that are more segregated by gender than those that don’t have any such linguistic distinctions? If so, we’d expect more egalitarian cultures to spring from entirely gender-neutral languages like Dari, the variant of Persian that’s spoken in Afghanistan. But as it turns out, the line between grammar and thought is simply not that direct. Languages have an enormous amount of leeway in expressing the same thoughts, and the specific methods they settle on are surprisingly arbitrary. That’s not to say that language scientists haven’t found any reliable influences of language on behavior. They have. But these tend to be fairly subtle. For instance, many languages force their speakers to sort inanimate objects seemingly randomly into grammatical genders—so to Spanish speakers, a chair is marked as feminine (la silla) but to German speakers, it’s masculine (der Stuhl). If you were to ask a Spanish speaker to imagine an animated chair as a cartoon character, he’d be more likely to choose a female voice for the character. But no one’s ever found a clear causal link between grammatical features of a language and the sort of large-scale societal behaviors that Chen argues for in his paper. And because Chen’s study simply looks for correlations, we can’t be sure that grammar is causing the behavior. It’s also possible that tense marking and live-for-the-day cultural attitudes spread together throughout populations without one causing the other. For example, I bet you’d find a correlation between tonal languages and the use of chopsticks at mealtimes, simply because both of these spread throughout a particular geographic region. But you’d be hard-pressed to tell a decent story about how the use of tone to distinguish word meanings leads to dexterity with certain dining implements. But still, we’re left with a puzzle: If language structure has quite a limited effect on the way we think and act, why then do we have these sturdy impressions that some languages are inherently more romantic, slovenly, logical, or fussy than others? The answer is that these impressions say less about the nature of those languages than they do about the strong associations we’ve forged between certain languages and the culture of their speakers. And these language-based associations can apparently trigger different behaviors. A particularly nice illustration comes from a study by Dirk Akkermans and colleagues, in which bilingual Dutch subjects played a business variant of the Prisoner’s Dilemma game, intended to test the degree of cooperative versus competitive behavior. (The game is set up so that you reap the highest profits if both you and your partner choose a cooperative strategy of keeping prices for your products high, and the lowest profits if you play cooperatively but your partner chooses to undersell you.) Half of the subjects played the game in English, and half played the game in Dutch—the idea being that the English language is more closely associated with highly individualistic and competitive cultures than Dutch. The subjects who played the game in English did indeed choose a more competitive strategy than those who played it in Dutch. But the effects of language on strategy choice really depended on how much direct exposure to Anglophone culture the subjects had. Among subjects who’d lived in an Anglophone country for at least three months, those who played the game in Dutch played cooperatively 51% of the time, while those who played it in English did so only 37% of the time. In contrast, among those who hadn’t spent more than three months in an Anglophone country, the rates for cooperative behavior were 48% for Dutch, and 45% for English. Actual proficiency in English had no discernible impact. So it’s not that English has any specific grammatical forms or even specific words that steer behavior in a competitive direction—it’s that English speakers tend to subscribe to more competitive norms of behavior, norms that the Dutch subjects subconsciously adopted when speaking English. The researchers may well have gotten very similar results if, instead of varying the languages, they’d exposed subjects to national symbols such as American versus Dutch flags, or pictures of bald eagles versus tulips. So in the end, perhaps learning to speak Mandarin would make Greeks somewhat more inclined to save for a rainy day. On the other hand, they just might get the same results by making a habit of eating with chopsticks.