Bryan Caplan  

The Degree and Origin of Foreign Language Competence

The Numbers Speak: Foreign Lan... Price Discrimination Explains ...
If you're curious about the underlying numbers for my last post, here they are.  The table shows every logically possible combination of (a) how well people speak a foreign language and (b) where they learned the foreign language.  Percentages should and do sum to 100%.

Table: The Degree and Origin of Foreign Language Competence



Where You Learned It








How Well You Speak It

Very well










Not Well






Hardly At All










Source: GSS, 2000 and 2006

[Note: "Didn't" doesn't mean that the respondent never studied a foreign language in school.  "Didn't" is what I assign to all respondents who say they don't speak a foreign language.]

Fun fact: While the most common status by far is "don't speak a foreign language/ didn't learn it anywhere," the second most common status by far is "speak a foreign language very well/ learned it in the childhood home." 

While I was reviewing these numbers, I recalled last year's debate with Tyler about the effect of upbringing on language.  I suspect he'll treat my table as vindication.  I disagree.  The data is more consistent with my original position that "You can make your kid semi-fluent in another language with a lot of effort." 

Few Americans are fluent in all of the languages their great-grandparents spoke.  The reason is clear: The fraction of people who learn a foreign language in the home is considerably smaller than the fraction of people raised by one or more parents who knew a foreign language.  Why do parents allow their ancestral tongues to fade from memory?  Because linguistic atrophy is the path of least resistance.  To get your kids in the "speak a foreign language very well/ learned it in the childhood home" box, you typically need to speak to them in that foreign language almost exclusively.  Unless you strongly prefer to speak that foreign language, that's a heavy burden.

Comments and Sharing

COMMENTS (8 to date)
Scott writes:

I feel language learning is dependent to a large extent on culture. In Switzerland, nearly everyone speaks at least 2 languages and it's perfectly normal for people to speak 4 or 5. In Spain, for example most people can't really speak English and in the US (and all other English-speaking countries for that matter) it's a miracle to find someone who can form a sentence in a foreign language.

Lee Kelly writes:


It may be partly to do with culture, but it's mostly to do with the costs and benefits. You said it yourself: it's a miracle to find anyone who can form a sentence in a foreign language in any predominantly English-speaking country. While English-speaking countries tend to share a cultural heritage, they are also quite different in many respects. Where they do not differ so much is in their aversion to learning foreign languages, and it's mostly because it involves lots of effort for very little reward.

Scott writes:

Spain and Italy would benefit greatly from an economic point of view if their citizens had an English competency rate as high as that of Scandinavia, Portugal, Germany or Austria. Even so, English competency remains comparatively very low in these countries.

Foobarista writes:

One question: what do they do with a Vietnamese kid who took Spanish in HS?

In other words, does the survey differentiate between "book-learned" foreign languages versus languages spoken at home?

Lauren writes:

Hi, Bryan.

I doubt it would substantively change your results, but nevertheless I think it might be informative to control for age if you do run any regressions from which you hope to draw more precise conclusions.

I speak several languages in addition to English, two of which I learned in school. In the first few years after taking my last course in each one, I'd have reported myself as speaking "well." But, even 5 years afterwards, I'd probably have reported myself as speaking "not well." My guess is that when you learn a language in school or via travel but don't practice it regularly--which practice is nearly impossible in America--skill deteriorates quickly. Age is a reasonable proxy for practice when it comes to Americans, particularly those who learned a foreign language in school. Anyone over 30 who learned a foreign language only in school is unlikely to retain it unless, say, the person's job involves using the language.

Although I agree with your hypotheses that language can be taught only on average at great expense (and I certainly don't think foreign language learning should be required by public schools to earn a degree), someone who disagreed with your hypothesis might credibly argue that exposing students to foreign languages in school is a way of helping students sort themselves into their comparative advantages, similar to exposure to music, sports, art, science, etc. I suppose that teaching engineering and science is also done only at great expense, and acuity is also only acquired or retained by a few. I don't suppose the survey asked similar questions about, say, physics or engineering, to the effect of: Did you learn physics/engineering/science in school/in the home/other? Do you do physics/engineering/science well/...?

Two disclaimers: First, I learn languages easily. I do notice, though, that my retention of grammatically correct speaking in non-native languages is dramatically lower compared to, say, my retention of math, engineering, physics, music, or the rules of English grammar. Second, I run a small not-for-profit that teaches English as a Second Language. That may cut both ways in terms of my biases.

Robin Hanson writes:

It seems odd that the table entries don't add up to 100%.

Alan Shields writes:

Would it be worthwhile to compare language mastery-retention vs algebra mastery-retention? It could be that we're really bad at teaching skills - even core skills like algebra.

There's also a reporting issue: my Spanish isn't bad. I can usually understand large portions of Spanish and make myself understood. With a little practice I can write at the early high school level. That said, when I was looking at the question my first answer was "Not Well".

This is because I compare my second-language knowledge of Spanish to others' first-language knowledge of Spanish. Indeed, compared to a native speaker my Spanish is awful.

Algebra might be less subject to this issue, though there's always the "I hate math" problem.

soonerliberty writes:

I have to disagree with the idea that it takes significant effort to teach a child a second language or that its language acquisition has to be exclusively in that language. True bilinguals or even polyglots expend almost no effort learning two languages or more at home. The only requirement is that they have input. My wife is Ukrainian (speaks Ukr, Russian, German, English) and I'm American (speak English, Russian, German, learning Spanish and Ukr). I speak English, my wife Ukrainian with our daughters. There is no effort involved. Children have an unlimited capacity to learn any language. Effort is not even relevant in a natural process. I think languages get neglected because people do not understand linguistics or how the brain acquires languages. Many doctors tell parents learning two languages simultaneously confuses the child or slows down language acquisition, but these ideas border on outright lies. There is no evidence of such a phenomenon. In fact, all evidence points the other way. Embarrassment is another factor. Parents may not want to be seen as foreign. They want to assimilate. That is certainly a cultural factor, but none of that has anything to do with effort.

Comments for this entry have been closed
Return to top