Full text of Professor Keith Chen’s talk: The Impact of Language on Economic Behavior at TEDxYale conference. In this talk, behavioral economist Keith Chen focuses on how people’s economic choices and attitudes are influenced by their language. In simple terms, does your language affect your ability to save money?
Keith Chen – Behavioral economist
Thank you very much.
So in the spirit of this afternoon’s conference, I want to take this opportunity to talk with you a little bit about some new work I’m doing and how it all started — because I saw a map that really freaked me out.
Let me show you that map and give you a little bit of background.
Just to start things off and explain, I’m a behavioral economist here at Yale. And so, one of the things I study is how people make decisions over time:
So, how people think about the future and how people think about the future that influences their behavior with respect to saving, with respect to studying for your exams, with respect to sticking to a diet, with respect to quitting smoking.
Now, what about this map freaked me out in particular?
So let me just say this was a map that was released by the European Science Foundation in the late 1990s. And in particular what freaked me out was this area in blue. Let me put it on a different map so it’s a little bit easier to recognize.
What this is is a map of Northern Europe. And what really, really kind of threw me for a curve was that the European Science Foundation had released a report that a number of kind of very, very reputable researchers had claimed that all of the areas inside this blue region were utterly and totally futureless.
That’s something of an extreme statement.
So I mean, as an economist, I’m someone used to, for example, making predictions that go horribly awry, but this almost takes the cake. You know, perhaps with the exception of Iceland – you can think about current European financial crisis – and the other areas inside this blue area actually are almost perfectly the countries that are doing the best.
And as an economist, what I would say is it seems crazy to call these places futureless. I mean, these places are all full of countries that are saving a tremendous amount of money, households that are saving a tremendous amount of money, countries that don’t have problems with their bonds and are investing a tremendous amount in public infrastructure and in the future — they just seem to care a lot about the future.
What I realized that led to this confusion, though, is that the team of researchers at the European Science Foundation led by kind of a superstar named Austin Dole, they weren’t talking about what an economist would mean when they say that a place is futureless, because this was a team of linguists.
And what they were saying was that, in fact, not that the households in this region don’t kind of care about the future, but that the languages in this region don’t really talk about the future in the same way that languages outside this area talk.
And what that led me to think about — which I’m going to tell you a little bit about right now — is the connection between economics, how you feel about the future and how your language forces you to talk about the future.
Okay. Let me explain a little bit about what that means.
So, for example – you can probably tell – I’m Chinese, and, you know, growing up I realized that Chinese families are different in many interesting ways.
What’s a little bit subtle — and that I didn’t realize till much later — is that the Chinese language actually forces Chinese speakers to talk about families in subtly different ways. So let me give you an example.
Suppose that a bunch of your friends come to you and say, “Would you like to go out for dinner?”
If you were speaking English with your friends, you could say, “You know, that sounds great. I’m really, really sorry, though, I have an uncle in town, and tomorrow I’m going to go out to dinner with him.”
Now, if you were speaking Chinese to your friends instead, actually, Chinese the language would force you to include a lot more information that I didn’t just say.
So, for example, there is no general word for uncle in Chinese; instead, what you’d have to specify, what you’d be forced by your language to tell your friends is whether this was an uncle on your mother’s side of the family or your father’s side of the family, and in fact, you’d be forced to say whether or not this was an uncle by birth or by marriage.
So, this is actually a very, very fundamental characteristic of language.
And as you see up there, the linguist Roman Jakobson expressed this best when he said, “Languages differ essentially in what they must convey and not in what they may convey.”
So in this sense, Chinese is forcing you to say a lot to your friends about the structure of your family in ways that if you’re an English speaker, you could very well think, “Well, they don’t need to know,” or “It’s none of their business.”
Now, let’s get back to Austin Dole and these kind of European linguists.
What these linguists at the European Science Foundation discovered was that when they looked at languages across the globe, a lot of global languages, what they discovered was that languages differ in a very, very fundamental way in the ways they force their speakers to talk about the future.
And they broke down languages into two rough categorizations: One, I’ll call them “weak-FTR,” or “weak future time reference languages” – are languages like Chinese, Finnish and German, which don’t force speakers, in fact, which allow speakers to speak about the future basically as if it’s the present.
And then other languages, like English, Greek, Italian and Russian – we’ll call them “strong-FTR languages” – force speakers to grammatically realize or to speak as if the future is something viscerally different than the present.
So for example, when I was just telling you how I would talk to my friends about taking my uncle out for dinner, if I were to say that in Chinese, it would be very, very kind of common and very, very easy for me to just say, ”I can’t go out to dinner tomorrow. I eat with uncle,” whereas to an English speaker, that just sounds very, very strange.
Now, a lot of people when I first showed them this list think it’s very strange because many people in this room probably know that English is a Germanic language and that English and German are close cousins, and yet as you can see, English and German find themselves on opposite ends of this divide.
Many of you probably speak German. One way you can think about this difference is, for example, suppose I was going to try and predict precipitation for tomorrow. In German, I could very easily say “Es regnet morgen” or “Morgen regnet es” or “Morgen ist es kalt,” and that sounds weird to an English speaker because what I’m literally saying is “Morning is cold” or “Tomorrow it rain” instead of “Tomorrow it will rain.”
Now, can this have an effect on your behavior? Can this have an effect on your economics?
Well, I did what, you know, economists with a crazy idea would do, and that is quickly try and dispel myself of crazy ideas by going and looking for as much data as possible around the world and trying to hit it as hardly as I can.
And let me just summarize the hypothesis that I was testing for you… that is, can languages that lead speakers to talk similarly about the present and the future lead those same speakers to feel similarly about the present and the future?
Now, why might that be important? Because if that’s true, then speakers of those languages should have an easier time saving, should have an easier time studying for exams, should have an easier time kind of not overeating and, for example, should have an easier time quitting smoking.
Let me just say, as a broad overview, that’s basically what I find. All of those patterns I just told you I find in spades in every major region of the world, and no matter how hard you try and hit this data and make it go away, you can’t get this pattern to disappear.
Let’s work through it a bit.
These are OECD countries that I put up in front of you. What does that mean? Well, these are generally rich kind of first world countries. They tend to have open markets and be liberal democracies.
We were talking a little bit about the European financial crisis. You can look all the way over there on the right — this is the average savings rate of countries over the last 25 years — and all the way over there on the right is Greece, okay? Saving just a little bit over 10% of their GDP, okay?
So, you know, that’s not such a surprise; we know they’ve had a problem with savings. It’s a little bit impolite to mention it in this audience, but if you noticed, we’re the United States, and we’re next in line.
Now, what I want you to notice, though — because I’ve colored a number of these bars in light blue — those light blue bars are those countries which speak languages that don’t make a strong distinction between the present and the future, okay?
And according to our hypothesis, that should make it easier to care about the future and easier to save. What we see is that’s very true.
Now, is this only a feature of rich countries?
Is this only a feature of well-developed economies?
No, here’s a much larger set of countries from all over the world. And what you see this kind of downwardly sloping line indicates is that exact same pattern seems to hold, you know, basically in every major region of the world.
If you speak a language that doesn’t distinguish strongly between the future and the present, you just save a lot more, okay?
Now, something else that this graph, well, can show you is something which provides an opportunity to hit this question much, much more hard.
And what is that?
Well, that is all of these countries, these seven countries you see in the middle of the screen, these are countries with multiple national languages.
And what’s fortunate about that is in many of these countries, they have multiple national languages and you can literally try and find families who live basically next door to each other in these countries but who speak different languages.
What is that going to allow us to do?
That’s going to allow us to look inside countries like Switzerland, where you see people who speak German, you see people who speak French, you see people who speak Italian, and you see families that speak Romansh, and countries in totally different parts of the world, like Nigeria, where you’ll find Hausa speakers living right next to Yoruba speakers living right next to Igbo speakers, okay?
What am I going to do?
Well, these are these kinds of – no, add up won more. You’ve got eight countries around the world that have this ability. And what I’m going to try and do is do what epidemiologists do and find matched pairs of families.
What does that mean?
Well, you could imagine — suppose I was standing up here on stage with 1.4 billion buckets, alright? And I took each one of you, and I sorted you into these buckets. Based on what? Well, based on the country that your family was born in and is living in; the sex and age of the head of the household; income – the exact income of your household; level of education; marital status –
It turns out, in Europe, there’re six different ways to be married; the number of children you find in your household; and finally and most powerfully, what religion you belong to — so 72 different world religions.
Now, that’s a lot of buckets — 1.4 billion. So, if you’re lucky enough, you might find yourself not alone but in a bucket with, say, one other family. That might be lucky for you.
You have a lot to talk about, you have a lot in common with this family. Lucky for me as a researcher, every now and then, two families find themselves in the same bucket, but they speak languages that treat the future differently.
So everything I’m going to tell you from here and now is true even when only comparing those families that are basically on every other dimension identical.
What do we see? What we see is exactly what we predicted, even after you hit the data with that 1.4 billion buckets. What we see is that households that speak languages that make a very, very weak distinction between the present and the future are 30% more likely to save in any given year.
Remember that’s already holding their income constant. They’re going to, by the time they retire, have accumulated 25% more wealth. They’re going to be 24% less likely to report having smoked intensively. That’s like more than a packet a day for a year at any given point in their lives.
And, you know, not just kind of monetary behaviors, but think about health behaviors: they’re going to be 30% less likely to be medically obese, they’re going to be 24% less likely to have smoked, and on almost every dimension, they’re going to be in measurably better health in the long run.
So grip strength, long capacity, walking speed, all of these measures, you could imagine your cumulative ability to kind of care about your future self — eat better, exercise and restrain from smoking.
All of those things seem to add up even when comparing families in the same bucket.
I’d like to leave you with this:
First of all, thank you very much for listening.
Second of all, this is research that’s really only just getting off the ground. Right now, a team of linguists, me, an economist, and a number of psychologists here in Yale are running experiments to try and identify exactly the psychological mechanisms by which these kinds of relationships are working.
And I invite you all to come to my website and kind of keep up to date with what I think is a really exciting new project, investigating what economists have to learn from linguists.
Thank you very much.