I've been reading Less Wrong for a while now, and have recently been casting about for suitable topics to write on.  I've decided to break the ice now with an essay on what living and working abroad in Korea has taught me which carries over into studying rationality.  While more personal than technical, this inaugural post contains generalizable lessons that I think will be of interest to anyone trying to improve their thinking.  

You may be skeptical, so let me briefly make my case that traveling offers something to the aspiring rationalist.  Many have written about the benefits of traveling, but for our purposes here is what matters:

Being abroad can make certain important concepts in rationality a part of you in ways studying can't match.

It's easy to read -- and to really believe -- that the map is not the territory, say, without it changing how you actually act. Information often gathers dust on the shelves in your frontal lobe without ever making it into the largely unconscious bits of your brain where so much of your deciding takes place.  

With this in mind travel can be seen as part of the class of efforts to learn rationality without directly studying the science, instead doing something like playing Go or poker, for example.   I don't know for sure, but such efforts could hold the promise of teaching us to incorporate insights into emotional attachment, statistical probabilities, strategy, maximizing utility, and the like -- things we've known for a long time -- into our instincts, deep down where they can actually change how we behave.  

I say all this because what living in a foreign country has given me is not so much a software update which has remade me into a paragon of rationality, but rather a hearty appreciation for certain facts which might make my thought-improvement efforts more fruitful.  No doubt many of you have already long-ago internalized all of this, and for you I won't be saying anything very profound.  

Nevertheless, here is what I've learned:

 

1) You are vastly more complicated than you think you are.  


The proposal for the Dartmouth conference of 1956, considered by some to be the birth of the field of AI research, had this to say:

 

An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.

 

Not to deny that considerable progress has been made in the past half century, but I think we can all agree that this thinking was just a tad bit optimistic.  

I'm not an expert on AI research history, but it seems reasonable to assume that these proto-AI researchers perhaps didn't appreciate how complex humans are.  You look at a triangle and you see a triangle; you reach for a coffee cup and grasp it; you start speaking a sentence and finish it with only the occasional pause.  What could be simpler?  We all forget our car keys sometimes, and some of us know a little bit about bizarre neurological problems like aphasia, but still.  In general we function so well that it never occurs to us that the things we do might actually be difficult to implement.    

The problem runs deeper than this, though, because there doesn't seem to be much in the way of techniques for elucidating this complexity from the inside.  If there were, neuroscience might've been discovered a millennium ago in East Asia by Buddhist adepts.  But instead our efforts at aiming the introspective flashlights on the machinery of our minds are thwarted by their presence totally outside our conscious awareness.    

Well, if you ever feel like you're not fully appreciating the intricacies of your wetware, sit in a coffee shop or bus stop in a foreign country while eavesdropping on people whose effortless bantering could not be more inscrutable, and you'll have it impressed upon youAlternatively, try to explain to someone with little-to-no English knowledge what something like "simple" or "almost all of" means.   Even without a bit of neuroscience training you'll start to get a grasp on the vastness of the gears and levers that make every utterance possible.  

This insight, at least for me, seems to creep into the rest of your thinking life, though in my case it's hard to tell because I've always pondered things like this.  It isn't a far leap from here to see the potential value of research into topics like Friendly AI. If human language and vision are complicated, what are the chances that human value systems are simple?  If you didn't manage to notice your retinal blind spot or the mechanisms by which you conjugate verbs in your native tongue, what are the chances that you aren't at least a little mistaken about your true goals and desires and how best to achieve them?  Exactly.  So maybe it's time to start reading those sequences, eh?  

 

2) Don't be bewitched by words

 

Obviously if you go to a country where English or a different language you're already fluent in is spoken, this won't apply as much. But my experience has shown me that living in and learning a foreign language bestows several valuable insights on those intrepid enough to stick with it.  Simply put, a sufficiently reflective and intelligent person could independently figure out about half of the sequence A Human's Guide to Words just by being in a foreign country and thinking about the experience.  

First you'd have to go through the shocking revelation that so much of what you say is a fairly arbitrary set of language conventions, and then you'd begin to relearn how to communicate.  You'd come to realize that words are mental paintbrush handles with which you guide the attention of other humans to certain clusters in thingspace, and that they are often disgusied queries with hidden connotations.  This will be triply reinforced by the fact that you'd often have to resort to empiricism to get your point across - accompanying the word 'red' or 'chair' by actually point to red things or chairs.  If you're spending time with natives the inverse will happen, and they will have to point to the parts of the world that words represent to communicate. You'll have a head start in replacing the symbol with the substance because you'll be playing taboo with nearly every word you know.  Since you'll be doing this with low-level language, it'll require elbow grease to port this into your native tongue when discussing topics like free will.  But if you can avoid slipping into cached thoughts, the training you received when you were a foreigner will likely prove useful.   

Beyond this, however, is the tantalizing possibility that we may be more rational when we think in a foreign language, perhaps because it increases reliance on the slow, analytic System 2 at the expense of the rapid-fire, emotional System 1. Psychologists from the University of Chicago tested this idea using English speakers proficient in Japanese, Korean speakers proficient in English, and English speakers proficient in French (Keysar, hayakawa, & An, 2011) [NOTE: I'm aware this study has been mentioned before on Less Wrong, but I believe this is the first actual discussion of the experiment and its methodology]. In the first few experiments participants were randomly sorted into two groups, one of which was given a test in their native language and one of which was given a test in the foreign language.  These tests were designed to elicit a well-known tendency for humans to differ in their risk preference depending on how the situation is framed.  

Here's how it works: imagine that you turn on the news today to find out that an exotic new disease is ravaging Asia, with an expected final death toll of 600,000.  The governments of the world decided that the best solution would be to design two separate drugs, and then to randomly select one reader of Less Wrong to decide between the two.  Your number came up, and now you have a choice to make.  

Drug A is guaranteed to save 200,000 people.  Drug B has a 33% chance of saving everyone and a 66% chance of saving no one.  

This is called the gain-framing, because what's emphasized is how many lives you'll save, or gain.  When framed this way, people often prefer to administer Drug A.  But studies find that if the same problem is loss-framed - that is, with drug A it is guaranteed that 400,000 people die while with Drug B there is a 33% chance that no one will die and a 66% that everyone will - far fewer people prefer Drug A, even though the results of using the drugs are identical.

Besides being sorted by foreign language participants were also randomly sorted by whether or not they got the gain or loss framing. Participants tested in their native language showed the predicted bias, but when tested in the foreign language, about an equal number of people preferred Drug A and Drug B.

An additional study found the same effect of foreign language on reasoning, but using a different bias.  People tend to be loss averse, preferring to avoid a loss more than they prefer to gain an identical (or slightly better) amount.  This means that people will often turn down an even bet which holds the possibility of gaining $12 and the possibility of losing $10, even though this bet has positive expected value.  As with the other studies, Korean speakers proficient in English more often showed this tendency when reasoning in their native language than when reasoning in a foreign one, especially for larger bets.  

There are a million reasons to learn a foreign language, but it'd be a very costly way to improve rationality.  With that said, for anyone willing to invest the time and effort, better thinking could be the outcome.  But even if you don't go to the trouble, simply trying to communicate with people who don't speak the same language as you will teach you a lot about how cognition and communication work.  

 

3) The Zen of the Unfamiliar


Living in another culture can make you aware of so many things that you previously failed to notice at all.  I remember not long after I got to Korea, I was in my kitchen and noticed that my sink was different from any of the ones I'd seen back in the States.  It was a single open pit sunk into the counter, with a strange spinning mechanism where the drain usually is.  After investigating for a while, I realized two things: one, the spinning mechanism was actually a multi-part contraption meant to catch food before it went down the drain (no idea why it could spin) and two, I'd just spend 100 times longer thinking about sinks than I had in the rest of my life combined.  

To successfully live in a foreign country you'll have to master the art of noticing things fairly quickly.  You'll start to watch how people dress, how they talk, how close they stand to each other, the relative frequency of eye contact, how they chew their food, what order people get served drinks.  You'll learn to read the environment to learn where to stand in line, where to catch the bus, where and how to buy things, which door is the exit and which one the entrance, whether or not certain places are likely to be safe, etc.  

You'll accomplish most of this by gathering evidence, forming hypotheses, using induction and deduction, and updating on new evidence.  The things you've been reading about on Less Wrong will be put to use in finding food and shelter, the tools of rationality will be your compass in a world where you can't read what's written on signs or buildings and most people can't understand your questions.  So there's a box on your wall with three buttons, two dials, a bunch of lights, and you're pretty sure it can make hot water come out of the shower?  Not a word of English anywhere on it, you say?  Well then you'll have to change one variable at a time and take note of the results, like any good scientist would.  

Being immersed in a set of shared cultural and linguistic norms that you don't understand makes almost every aspect of your life an experiment.  It's exhausting, and one of the most informative experiences I've ever had.  On an emotional level, it will teach you to be more at ease with partial understanding, frustration, and confusion.  With your comfort zone an ocean away, you'll either persevere and think on your feet, or you'll end up sleeping in the rain.  

 

__

 

Like with learning a foreign language, there are many reasons to travel abroad and experience another culture.  And of course, a plane ticket alone is not enough to make you a better thinker.  But if you know what to look for and are actively seeking to grow from the experience, I can attest that being foreign for a little while is one way to become a bit more sane.  

 

 

 

 

New Comment
42 comments, sorted by Click to highlight new comments since: Today at 3:02 AM

Well done!

A nitpick:

you'll have to change one variable at a time and take note of the results, like any good scientist would.

In the shower control problem, this is a satisficing approach. Any good scientist, however, ought to know that changing one variable at a time is incredibly wasteful relative to fractional factorial experimental designs, which are themselves not optimal designs in general.

Alternatively, try to explain to someone with little-to-no English knowledge what something like "simple" or "almost all of" means.

I remember having this conversation with someone with very poor English:

Me: Did you have fun tonight?

Her: Fun? What does that mean?

Me: Did you enjoy yourself?

Her: Move myself?

Me: Enjoy.

Her: [confused look]

Me: Never mind.

"Did you smile?" /smiles

"Oh!"

I imagined myself explaining “simple” and the very first thought was the phrase “not complex”. Oops.

Just show them examples of small Turing machines until they catch on.

"Alright kids, can you say 'computation'? Repeat after me: COM-PYOU-TAY-SHON..."

"Foreign" does not mean "stupid".

Just because you're not stupid doesn't mean it's trivially easy for you to correctly pronounce a word in a foreign language after hearing it once spoken at a normal speech rate.

As for the “kids”...

As a person who had to adjust to life in the United States after moving over from Russia... There are three English phrases that a foreigner must know to learn English quickly, so long as they are willing to LOOK stupid (an important art in a world so obsessed with being serious).

  1. "Where's the Bathroom?" Apart from its obvious uses, it is is essential to one's survival to know where one may hide to plot one's next move. Given the creative responses I sometimes received, I suspect it is also useful for learning profanity.
  2. "I don't understand." It is the most useful phrase in the English language. People will generally make an effort to communicate on a number of levels after this is used. Excellent for new words and concepts.
  3. "Please help." Not all technology will have clear instructions printed on it. Getting lost is also unpleasant when you cannot read the signs. So, asking for help early and often helps one avoid all kinds of trouble (such as accidentally setting a classroom on fire because you don't know how a Bunsen burner works). Many learning opportunities are missed when one figures something out with no input from a person used to the tech...

Perhaps phrases similar to these exist in every language--by using them, one can learn the rest of the language quickly, if clumsily.

Math works well with the kids. I write "1 + 1 = 2" and say "simple", then write the quadratic equation and say "not simple".

Or, if one of the kids is Eliezer Yudkowsky, you can write Maxwell's equations and say "simple", then write a program simulating Thor and say "not simple".

So, simple is a type of equation?

Jokes aside, that's an excellent answer.

I see nothing wrong with this. You may have to go through complex to get to simple.

If you didn't manage to notice your retinal blind spot or the mechanisms by which you conjugate verbs in your native tongue, what are the chances that you aren't at least a little mistaken about your true goals and desires and how best to achieve them?

Even though I'm very familiar and comfortable with your thesis, I found that sentence striking.

Striking good or striking bad?

"There are a million reasons to learn a foreign language, but it'd be a very costly way to improve rationality."

It is a "free" side-effect if you belong to the 95% of the world population without English as native language.

It sounds like most of these experiences could be had by interacting with foreigners in your own country, too. Of course immersion helps.

You could get quite a bit of the way there, yeah. Interestingly, it's possible to have the inverse experience: lots of foreigners living here make almost no effort to interact with locals, learn the language, or experience the culture in any significant way.

I doubt there's anything quite like full-on immersion in a culture you know nothing about, though.

There's a clear difference between interacting with foreigners in your own country, and being a foreigner in another country, which is basically that when you're a foreigner, it's your beliefs/customs/mannerism that are being questioned. If everybody faces the back of the elevator, you're going to start pondering why in the U.S. you face the front of the elevator, whereas those facing the back wouldn't stop and think about why a foreigner might be facing the opposite direction.

Also, obviously, the immersion factor is there. Speaking to foreigners in your own country is not nearly as new/scary an environment as being in a completely foreign country.

Is there really a why to the direction of facing? Or is it just that both states are stable, but which one is chosen is due to uninteresting historical contingency?

Facing the front of the elevator seems to be the better choice - you can press buttons, watch the floor numbers tick up, and exit the elevator more quickly.

Facing backwards also sort of cuts out human interaction in cases where, say, somebody new enters the elevator.

The buttons are (typically) on the front of the elevator. If they were on the back, you would have to move further to press them, and it would be harder to press them in an elevator that already had some people in it. Given buttons-at-the-front design, in order for everyone to face the back of the elevator, they would have to press the button for their floor, and then turn around, and then, when their floor is reached, turn around again.

This article is really good. The title isn't, but I think it enhanced my overall experience by making me go in with low expectations. Nice work!

Not exactly what I was going for, but I'm glad to see it had a net positive effect. Incidentally, what do you think was wrong with the title, and what would you have called this article? I originally had it titled "what living abroad can teach a rationalist", something descriptive, but I thought this might catch people's attention. I struggle with coming up with good titles for papers and sections.

I thought it was going to be about something like this.

"what living abroad can teach a rationalist"

Huh, please no; "... can teach you" would be fine.

Thanks for the advice. Incidentally, does anyone know what the policy is for cross-posting things you write for LW? I looked but couldn't find anything. I have a personal blog where this essay would be at home, and I'd like to cross-post so that the blog can function as a kind of resume and repository for my writing.

My sense of the norms is that cross-posting is fine. Examples of cross-posting: one, two, three; discussions where cross-posting is mentioned favorably: one, two.

Cross-posting would be UNFORGIVEABLE.

No, I lie, I don't even know why that would be a problem.

I am kind-of bothered when there are two different comment threads about the same article in two different places, but I think this is a pet peeve of mine shared by hardly anyone else.

I'm not crazy about the same post at multiple locations because if I'm interested in the discussion, I have to follow it in two places.

I don't know what I would have called the article, but generally I don't think using the word "sane" is optimal. I honestly thought this might be something to do with culture wars/reactionary politics at first glance.

The title lead me to expect an article about seeing through peoples' foreignness for the person they really are. Going from thinking "they do things in that silly manner because their culture has mislead him and made them insane," to understanding that they've probably never had the chance to question their culture's ways and they would if you'd give them one.

As an exploration, it would have been of great value to me and, I believe, to the rest of the lesswrong community. We are strange. At some point we're going to have to start wearing that on our sleeves to be true to ourselves, but we need outsiders to be able to see through our strange practices, to see that we're not a big crazy cult. If we knew what we'd have to do to look past others' foreignisms, we'd be better equipped to lead them to do the same for us.

P.S. Please write this article.

Cool. I was going to write a post in which I suggested a mental exercise involving pretending that words don't have meaning, and expected that the exercise would be very easy to do for people living in a country foreign to them. But this is better than that post would've been because you actually did it!

Good to hear! Yeah, spend any significant time abroad and you'll start to replace the symbol with the substance, trust me :)

'n Mens kan defnitief 'n bietjie van tweetaaligheid leer. Woorde wat dieselfde klink in een taal kan baie anders klink in 'n ander taal - byvoorbeeld, 'skoon seun' en 'skoonseun' klink nooit nie dieselfde nie in Engels. Woordorde kan ook baie verander.

(For those who are curious, I've rot13'd the English translation of the above paragraph below)

N crefba pna qrsvavgryl yrnea na ovg sebz ovyvathnyvfz. Jbeqf gung fbhaq gur fnzr va bar ynathntr pna fbhaq irel qvssrerag va nabgure ynathntr - sbe rknzcyr, 'pyrna fba' naq 'fba-va-ynj' qb abg fbhaq gur fnzr va Ratyvfu. Jbeq beqre znl nyfb punatr n ybg.

Alternatively, try to explain to someone with little-to-no English knowledge what something like "simple" or "almost all of" means.

It's easy to explain "almost all of": draw a picture of "all", then "none", then "almost all". The first time you do this, you'll expect it to work, but it won't make sense to the other person, so you'll have to do it again, differently. (I think my first attempt would be to draw a bunch of small things, then to show circles enclosing all, none, and almost all of the things.)

It's harder to draw "simple". Possibly if you draw some simple geometric shapes (straight line, circle, etc.) and some complex geometric shapes (curved line, several lines or circles, some complicated geometric shape, etc.) and then label some as "simple" and some as "not simple". Then start drawing comparisons ("more simple and "less simple") to explain that it's a scale. Make sure to falsify as many alternate hypotheses as you can think of (such as "one line" and "many lines").

And this is still much easier than it should be: a foreign language speaker already knows the concepts of "almost all" and "simple" under different names, and this saves you. The foreign language speaker has to do a lot of work to understand the explanation, too.

You quickly learn to not rely on words. Communication, especially with my students, involves lots of drawing, lots of acting, lots of examples, hence what I wrote about having to rely on empirical examples of things. That in itself is a pretty valuable experience I think.

You might be making teaching English harder than it is, since you don't really have to teach concepts that the students already know in their own language, you just have to translate them. A good English-Korean dictionary would go far. There are of course some terms that won't map over 1-to-1, but the majority should.

Simply put, a sufficiently reflective and intelligent person could independently figure out about half of the sequence A Human's Guide to Words just by being in a foreign country and thinking about the experience.

I always assumed a sufficiently reflective and intelligent person could independently figure out most of the Sequences just by living as a human.

Available evidence seems to point to the contrary, unless you are using a quite high value for "sufficiently", higher than the one used by fowlertm in the quoted phrase.

Well, I'm basing this on having seen comments talking about how people had already figured out large parts of the sequences, and my own experience of having independently come up with large portions under different names and being very pleased to find someone had already done so and was now building on them.

There's probably a degree of availability bias and so on, it was just my general impression.