When do people learn languages?

• People want students to learn foreign languages.
• People want immigrants to learn the national language.
• People want their home language to be the national language.
• Minorities are concerned that their languages not disappear.
• Esperantists want everyone to learn Esperanto.
• Non-Americans want to learn English.
• Non-Americans are worried that everyone will learn English.
• Parents may want to teach their children something besides the national language

My purpose with this page is not so much to argue for or against any of these goals, but to set out the facts (as far as they're known), and draw some mostly unwelcome lessons.

Languages are learned in school

These courses typically take 3 to 5 hours a week; one might expect more of bilingual or total-immersion schools. But even here the results are not perfect. Some Canadian schools teach French to anglophones by immersion; it's reported (François Grosjean, Life with Two Languages, p. 219) that the students do well enough academically, but have "great difficulty" communicating with French-speaking Canadians outside the school, and don't initiate conversations in French.

European schools have a better reputation than U.S. ones, but I'm not convinced that the situation is spectacularly different. After ten years of English, students may do quite well; but if they don't have ongoing opportunities to practice, their English may not last much more than the American students' French or Spanish.

The basic fallacy here is to take learning as an irreversible process. Because someone learned something in school, whether it's Latin or trigonometry or the exports of Venezuela, it doesn't mean that they still know it. An example is the father of a friend of mine, a Brazilian who studied in the US for a year. Forty years later, he can still read English, but he's quite unable to carry on a conversation in the language.

To sum up: school courses may help you to learn a language (see below); they can't be counted on to teach it to a whole population. Immersion schools will probably work, though even that will be of no use if the student doesn't keep using the language after graduating.

My ancestors learned by 'sink or swim'; so can you

First, if those ancestors immigrated as adults, then very likely they didn't learn the national language well, if at all. For Americans, at least, the details are generally several generations back and misty. We forget just how multilingual the country was at the height of European immigration. People generally moved into ethnic enclaves, married among themselves, and worked in menial jobs where there was little need for English. There were daily newspapers in half a dozen languages in all the major cities (indeed, a surprising number of them still exist). Much of the bureaucracy we have to deal with today didn't exist: income tax, health plans, credit card contracts, forms required by schools. Older immigrants could often get by with a minimal command of English, or none at all.

Children might indeed be sent to a school which taught in English only. This 'worked' in the sense that the child learned English; but this is far from proving that this is the only or best way to treat immigrant children. At the very least these children missed a year or two of instruction, and the later they came, the less command of formal English they would have attained.

And maybe that was acceptable in 1880 or 1920, when most jobs required little education and minimal formal English. Today almost all jobs require a high degree of schooling; and complicated forms and directions are inescapable. What worked several generations ago may be insufficient today.

I know this guy who speaks ten languages!

For such people, the brutal facts about most people's language learning-- i.e., that they don't do it-- may be hard to believe or sympathize with. We learned French, and enjoyed it! Look at all those Larousses on the shelf! We had so much fun talking to that schoolteacher in Versailles, buying French rap at FNAC, reading Daniel Pennac, answering that Tech Support call from Montréal ... can't people see that learning languages is both fun and useful?

Well, no, any more than people in general see the fun in manga, or algebra, or skeet- shooting, or ska. Like it or not, language geeks are a minority, and their abilities are no guide to language policy.

Children learn languages easily

Children begin learning languages at birth (infants pay attention to their parents' voices, as opposed to random noises or even other languages), and haven't really mastered it subtleties before the age of ten years. Indeed, we never really stop learning our language. (See David Singleton, Language Acquisition: The Age Factor, p. 56.) This isn't exactly the sort of behavior (like foals walking an hour after birth) that we call 'instinct' in animals.

But at least it's effortless, isn't it? Well, no, as we can see when children have a choice of languages to learn. What's found is that, to be frank, children don't learn a language if they can get away with not learning it.

Many an immigrant family in the U.S. intends to teach their child their native language; and for the first few years it goes swimmingly-- so much so that the parents worry that the child won't learn English. Then the child goes to school, picks up English, and within a few years the worry is reversed: the child still understands his parents, but responds in English. Eventually the parents may give up, and the home language becomes English. An anecdote from Grosjean:

Cyril, a little French boy in the States, started going to an English-language day care center, he brought home English-speaking friends, he watched television, and American friends of his parents quite often came to dinner. Above all, Cyril realized that his parents spoke quite good English, and as there was no other reason for speaking French (no French-speaking grandmother or playmate, no French -speaking social activities outside the home), Cyril probably decided that the price to maintain both languages... was too high. Little by little he started speaking English to his parents and ceased to be an active bilingual, although he retained the ability to understand his fist language.

The linguist R. Burling spent two years in the Garo Hills of India; his son grew up speaking English and Garo-- mostly the latter. They left India when the boy was three; for awhile the boy would attempt to speak Garo with anyone he met who looked Indian. But within six months, he wouldn't speak any Garo and seemed to have trouble with even simple Garo words.

A child is likely to end up as a fluent speaker of a language only if there are significant people in her life who speak it: a nanny who only speaks Spanish, a relative who doesn't speak English, etc. Once a child discovers that his parents understand English perfectly well, he's likely to give up on the home language, even in the face of strong disapproval from the parents.

These stories help demonstrate that it's a myth that children learn to speak mainly from their parents. They don't: they learn mostly from their peers. This is most easily seen among children of immigrants, whether they come from differing language backgrounds or merely different dialect areas: the children invariably come to speak the dialect of their neighborhood and school, not that of their parents. (I found a neat example of this in my college's alumni magazine: A liberal family in Mississippi sent their daughter to the public schools, which except for her were all black. She grew up speaking fluent African-American Vernacular English.)

Supporters of the 'language instinct' make much of the fact that children learn to speak without formal instruction-- indeed, they notoriously ignore explicit corrections. For instance, an example collected by Martin Braine [Pinker, p. 281]:

Child: Wnat other one spoon, Daddy.
Father: You mean, you want the other spoon.
Child: Yes, I want other one spoon, please, Daddy.
Father: Can you say "the other spoon"?
Child: Other... one... spoon.
Father: Say "other".
Child: Other.
Father: "Spoon."
Child: Spoon.
Father: "Other... spoon."
Child: Other... spoon. Now give me other one spoon.

The apparent effortlessness is largely an illusion caused by psychological distance. We just don't remember how hard it was to learn language. (In fact, there's some studies suggesting that memory is tied to language, so that we can't remember the language learning process.) The perception of effortlessness should be balanced, anyway, by the universal amusement (which some cartoonists have been mining for nearly half a century) over children's language mistakes.

Another anecdote: my wife liked to tease a young boy who was having trouble understanding the reciprocal nature of personal pronouns:

Adult: Is this your ball?
Child (trying to say it's his): Yes, it's your ball!
Adult: So, it's my ball?
Child: No! (cries)

That's not exactly easy!

Another clue that children find language difficult is that they become agitated when someone speaks the 'wrong' language. An English-German bilingual child, Danny, was speaking to a German-speaking researcher; trying to help, his mother (who normally only used English with him) asked, Was macht der Vogel? ("What's the bird doing?") Danny, startled, told his mother, Nicht 'Vogel'! ("Not Vogel!") He points to the researcher and said Du Vogel ("You bird"), and to his mother and said Du sag 'birdie' ("You say birdie").

Another example: an Italian-German bilingual girl, Lisa, became upset and started to cry when an Italian friend spoke to her in German. On another occasion, Lisa's father said something to her in German, and she responded, No, tu non puoi! ("No, you can't!") Keeping two largely unknown language systems separate is a tricky task, and associating each with different people helps: Lisa can count on knowing that whatever Daddy says is Italian. If anyone in her life could use either language at any time, the learning task would become much harder.

One may fall back on the position that language may be hard for children to learn, but at least they do it better than adults. This, however, turns out to be surprisingly difficult to prove. Singleton examined hundreds of studies, and found them resoundingly ambiguous. Quite a few studies, in fact, find that adult learners progress faster than children (Language Acquisition, pp. 94-106). Even in phonetics, sometimes the last stronghold of the kids-learn-free position, there are studies finding that adults are better at recognizing and producing foreign sounds.

Now, I think Singleton misses a key point in understanding this discrepancy: the studies he reviews compare children vs. adults who are learning languages. That's quite reasonable, and indeed it's hard to imagine an alternative approach; but the two groups are not really comparable! All children have to learn at least one language; but few adults do. So the studies compare the situation of all children with that of the minority of adults motivated to formally learn other languages.

Why do children learn languages well, when even adults who want to learn them have trouble with them? Innate abilities aside, children have a number of powerful advantages:

• They can devote almost their full time to it. Adults consider half an hour's study a day to be onerous.

• Their motivation is intense. Adults rarely have to spend much of their time in the company of people they need to talk to but can't; children can get very little of what they want without learning language(s).

• Their peers are nastier. Embarrassment is a prime motivating factor for human beings (I owe this insight to Marvin Minsky's The Society of Mind, but it was most memorably expressed by David Berlinski (in Black Mischief, p. 129), who noted that of all emotions, from rage to depression to first love, only embarrassment can recur, decades later, with its full original intensity). Dealing with a French waiter is nothing compared with the vicious reception in store for a child who speaks funny.

If adults could be placed in a similar situation, they might well learn languages as readily (I don't say 'easily'!) as children. The closest such situation I can think of is cross- cultural marriage. And indeed, this works quite well. My wife, for instance, a native Spanish speaker who came here in her late 20s, has learned exceptional English, since we speak it at home. By contrast, some of her Spanish-speaking friends of the same age, married to other Spanish speakers, speak English haltingly and with a strong accent.

Languages take immense effort to learn, and people will only learn them if it's socially or economically inescapable.

So what makes learning inescapable?

Again, we can learn from the children. Mere opportunity isn't enough. As we saw, immigrant children have an excellent opportunity to master their parents' language-- but, if possible, they don't. They learn only if they can't interact with relatives otherwise (e.g., if the relatives are monolingual).

Similarly, adults may be exposed to other languages-- their neighbors speak one, or their employees do, or it's on another TV channel, or it's taught at a nearby school-- but these sources are easily ignored.

The exposure has to be from people you need to interact with. For adults, this means principally employers, relatives by marriage, children, and clients or vendors who don't speak a language you already know.

The importance of these sources varies, of course, with time and quality. Very little is needed to learn to buy groceries in another language, or to learn how to do a menial job. Immigrants will probably not master the local language unless they work mostly with natives, or marry a local; their best chance will probably be to learn the language from their children.

Note that harrassing immigrants by requiring government forms to be in English will accomplish little. Bureaucracies are important and intimidating, but no one interacts with them enough that they force language learning. The usual strategy is to take a friend or child along to translate.

'Need' must be interpreted from the learner's perspective, not the observer's. I discovered this when a Peruvian couple (old friends of my wife's) stayed with us for a few months. They couldn't find work in Peru, so they certainly had good reason to look for work here and to learn English. But they approached both tasks haphazardly. The husband had a tutor for learning English-- but he discovered that the tutor spoke Spanish, and so they chattered away in Spanish. When he wasn't noodling with his computer, he looked for jobs-- mostly by checking the want ads in the local Hispanic paper. They ended up moving to Spain.

It's a bit like weight loss: people will say they "need to lose weight", but they'll believe any fad or scam rather than actually eat less or exercise more. Our Peruvian couple didn't take language learning seriously so long as they had a backup plan of moving to Spain.

Children growing up in a monolingual environment can't avoid the work, so they learn the language. Everyone else, if possible, doesn't.

The market for a language

A bit cynically, but realistically, I answer: the ensemble of those who find it completely unavoidable. No amount of rationalization, publicity, or moral suasion will increase the figure-- though it may increase the language's popularity among language geeks.

The "language problem"

Learning a language well enough to negotiate business deals, for instance, could easily cost an adult five years of time and $10,000 in tuition. An executive or bureaucrat doesn't have that sort of time-- nor is the capability worth the time investment. He hires interpreters and translators instead.

It's certainly a disadvantage that so much interesting material, from books to movies to music to comics, is published in languages we can't understand. But we can hire translators too, or rather let the publishers do so. Material in other languages is not so much unavailable as delayed.

Of course, not everything gets translated-- frustratingly so, if you know something about the source culture. Informed comics fans, for instance, are puzzled why Lewis Trondheim or Ralf König, enormously popular in Europe, and by no means difficult or inaccessible, are almost unknown in the U.S. However, there may be reasons for this. Americans aren't used to long-form comics-- they simply have no category to put Trondheim in, and vaguely compatible English-language works (say, Joe Matt or Dan Clowes) are almost as obscure. And Americans are not quite ready to read gay-themed books like König's, however funny.

By contrast, manga has made a spectacular conquest of American youth... when I see young cartoonists doodling, it's more likely to be an imitation of manga than of Disney or Peanuts or even superhero comics. And the desire for more manga and anime has led a surprising number of fans to try to learn Japanese.

The cost of learning a language

Such efforts are not so much wrong as misleading. First, they benefit from the specious impresiveness of large-scale sums. Heck, chewing gum is a $4 billion industry in the U.S. alone. In a $6 trillion economy, it's not hard to wallop people with high numbers.

The calculations should be done per capita. How much do I lose by not knowing Arabic? Then let's compare that to the cost of learning Arabic-- or, if you like, your favorite auxlang. This should be compared to the cost of hiring a translator for the times I need to talk to an Arabic speaker or read Arabic media. And let's not forget the opportunity costs: the money and time it takes to learn Arabic or Esperanto isn't available to do something else.

I suspect that if you did all these calculations, you'd find that we're already minimizing language costs, and that spending more money on teaching a rarely-used auxlang to everyone would be uneconomical. Translators are awfully cheap, considering.

Languages over time

Historically, a language's fortunes begin to rise when its speakers conquer a wide swath of territory. The need to learn a conqueror's language is almost always greater than the conqueror's need to learn that of the conquered; that's why the languages of Latin America are Spanish and Portuguese and not Quechua or Nahuatl, those of the Mediterranean are Romance and Arabic and not Berber or Hittite, those of China are Sinitic and not Thai or Zhuang. (Exceptions arise when the conquerors are few. E.g., when about 50,000 Normans conquered England, they were eventually swamped by the several million English; the same can be said of the nomadic conquests of China.)

We shouldn't exaggerate the interest in learning the conqueror's language. From what we see in current societies, we may well doubt that many Incas learned Spanish, or Gauls learned Latin, or Mesopotamians learned Arabic. In pre-modern societies, where schooling and religion happened entirely within the ethnic community, even their children may not have learned the invaders' language.

An interesting case study is found in Bruce Mannheim's The Language of the Inka since the European Invasion. Sometimes the Spanish complained that the Indians continued to speak Quechua... but in fact the conquerors found it very handy to dominate a people who couldn't speak their language, and thus couldn't protest or understand the laws used against them. Measures were sometimes taken against community leaders who learned Spanish too well.

Over time the conquerors' language does spread. Some of this is due to very basic reasons: the conquerors directly colonize the new territory; or they take wives from among the locals, resulting in a quick population boost. (English took over in North America due largely to the first factor; Spanish and Portuguese in Latin America due to the second.) Conquest also allows other outsiders entry, and they or their children will pick up the dominant language (this has reinforced English, Spanish, and Portuguese in the Americas). If people from different linguistic backgrounds mix, they're forced to use the national language-- which is why American blacks speak English and not any African languages.

Some number of outsiders will also learn the language, either to deal directly with the empire, or to work as translators. Scholars and others may learn the language for cultural reasons; this effect can persist long after the empire has decayed or fallen. Latin was still widely learned in Europe 1500 years after (the Latin-speaking half of) the Roman Empire disappeared.

In modern times, national languages are reinforced by universal education, as well as by conscription, economic mobility, and other forces that mix up populations. Modern media are often blamed for spreading language, but this is doubtful: people don't try to speak like radio announcers and TV actors; they try to speak like the other people in their neighborhood, school, or barracks.

As with anything that's difficult to master, languages can generate affection, and people are tempted to explain the cultural influence of a language by its inherent excellence. Thus the French like to talk about the clarity and logic of French. Pay no attention; literary and cultural influence ultimately depends only on conquest. Conquest makes nations big and rich; big populations produce more artists, and rich ones pay for more art. For all we know, Lappish is the world's greatest language to write poetry in, but we'll never know, because the Lapps neglected to ever conquer a bunch of their neighbors.

Eventually people abandon a language, a process called language death. This is a sad process, especially in the case of unwritten languages whose grammars, lexicons, quirks, and cultural resources are entirely lost; but it's also a natural one, and by no means limited to modern times. (As Latin thrived, it supplanted Oscan, Umbrian, Faliscan, Venetic, Etruscan, and other languages of pre-Roman Italy.)

Linguists have studied the process (for a report see David Crystal, Language Death), and found that it proceeds in stages, often over a couple of generations: the language is used in more and more restricted contexts, and its grammar simplifies, as its more obscure rules are lost. In Women, Fire, and Dangerous Things, George Lakoff considers Dyirbal, whose gender system provided his title. When the language was fully alive, its four genders had some rather complex rules: e.g. everything female belonged to the feminine gender, but also objects used mostly by women, and certain birds which were considered to have the spirits of female ancestors. As the language gave way to English, the rules were simplified, till the last speakers used the gender only for female humans.

(Thus, when you hear about the last three speakers of a language, all in their 90s, the language is probably already dead. If they're the last speakers, they may not have learned the grammar or lexicon in its original extent, or at best they've forgotten most of it.)

Advice for language learners

Given the discussion so far, the prospects for language learning may seem pretty bleak. It seems that you'll only learn a language if you really need to; but the fact that you haven't done so already is a pretty good indication that you don't really need to. How to break out of this paradox?

At the least, try to make the facts of language learning work for you, not against you. Exposure to the language, for instance, works in your favor. So create exposure.

• Read books in the target language.
• Better yet, read comics and magazines. (They're easier, more colloquial, and easier to incorporate into your weekly routine.)
• Buy music that's sung in it; play it while you're doing other things.
• Read websites and participate in newsgroups that use it.
• Play language tapes in your car. If you have none, make some for yourself.
• Hang out in the neighborhood where they speak it.
• Try it out with anyone you know who speaks it. If necessary, go make new friends.
• Seek out opportunities to work using the language.
• Babysit a child, or hire a sitter, who speaks the language.
• Take notes in your classes or at meetings in the language.
• Marry a speaker of the language. (Warning: marry someone patient: some people want you to know their language-- they don't want to teach it. Also, this strategy is tricky for multiple languages.)

And since motivation works in your favor, create motivation.

• For your reading, read things you're interested in. I think many people are turned off by languages because teachers foist works on them that they'd never read in their own language. Pop culture is often the best choice... there's a reason pop culture is popular: it's easy to digest. But if it's 17th century madrigals or tax law that turns you on, read those.

• Take a trip to the country where the language is spoken-- better yet, study there for a year. Don't hang out with Americans there; go talk to the locals. Panic is very motivating.

I've spoken of kids' brutality toward kids who speak funny as an incentive; but this one doesn't work as well for adults. Maybe this is biological-- as Marvin Minsky suggests, nature wants offspring to imitate their parents, not vice versa-- or maybe it's just that adults are better at avoiding hazing. When it comes to language learning, however, this means that many adults are terrified of saying something wrong, and just say nothing at all. Needless to say you don't advance very far this way.

People often complain that people "won't speak the language" with them. It may help to realize that this is a universal strategy of bilinguals: if you're speaking with someone, use the language that makes for the most effective communication (see Grosjean p. 135). I've often had the experience of starting out a conversation in French, going on swimmingly for awhile, then making one mistake, and the other person immediately switches to English.

This can be quite frustrating when you're looking for opportunities to practice; but you might want to remember that random foreigners are not paid to be language teachers. If you want more practice, just keep using the foreign language and let them respond in English. (This was actually standard procedure for the Apollo-Soyuz missions: the American astronauts spoke Russian and the Russian ones spoke English. This is an excellent compromise for a situation that demands clarity: it guarantees that no one's babbling away without being understood.)

A hint from cognitive science: our brains work by making contextual connections. This is one reason systems analysis is difficult: if you want a computer to do something that humans do now, you need to know about every contingency; but the people who know how to do it can't list every contingency for you. They remember them in context.

So, memorizing vocabulary lists is not likely to be effective, because the words will be remembered only in the context of reading vocabulary lists-- that is, not when you need them. Talking to people is better, because words will get associated with real-life situations.

The number of connections seem to matter, too. If you just remember "jaune means yellow", you'll probably quickly forget it. But if you say the word out loud, write it out, hear it on a tape, think about the soleil being jaune, and link it up with 'jaundice', your chances are better.

While we're at it, here's some tips from my own experience that I can't claim linguistic backing for:

• Concentrate on learning words, not on grammar. My favorite example: if a foreigner comes up to you and asks, "Train station where please?" you can understand and help, even though the grammar is awful; but you can't help a bit if he says, "Can you please tell me where to find... uh... er...oh, damn.. oh, the place where the... uh...." There's plenty of time later for learning the grammar; but knowing the words will help you communicate now, and help you in learning the grammar later.

• If you're learning on your own, don't just buy a book; buy tapes. You need to hear the language. Plus, tapes are great for playing in the car.

• Don't be unnerved when you can't understand music or movies-- these are among the hardest types of spoken language to follow. One-on-one conversation is a lot easier. You're guaranteed to understand half the conversation (your part), and your interlocutor will help with the rest. You can't easily ask a movie character to repeat himself.

• Use the dictionary-- but sparingly. Not many people can enjoy a book if they're looking up a word ten times a paragraph. See if you can get the gist without looking everything up. If you can't, try an easier book.

• Be playful. Make jokes, look for puns, don't be afraid of making a fool of yourself.

• As with weight-lifting, you can get to a plateau where you're not really improving. Look for something to increase the challenge: read slightly more difficult books; give those movies a try (the hardest Spanish I know is that of Cantinflas); take up crosswords.

• Evaluate the methods you're using, and start over if they're not working. Some people like traditional methods-- classes and textbooks-- but these bore other folks silly. Personally, I recommend reading comics-- they're fun, the language is more colloquial (thus a good preparation for meeting actual people rather than literary characters), and the pictures give enough context that you often don't need the dictionary.

Implications for language planning

Crystal's Language Death offers a number of suggestions, which I won't repeat here because the book is at the library; I'll just note that there's no guarantee. In the terms of this paper, you have to make people feel they need to learn the language.

As with individual learning, exposure helps. I suspect that local-content laws and signage laws have an effect. I doubt that (say) any Québec anglophone's decisions are changed by seeing signs in French, but an immigrant to the province might be subtly affected-- the signs increase the language's presence. Similarly, the Peruvian government once required an hour of Peruvian content a day on the radio; at the very least this got a lot of people familiar with traditional waynos, mostly sung in Quechua.

More important, perhaps, is to increase the domain where the language can be used. Language death involves a steady shrinkage in the areas where the language is spoken; and defeat is already in sight if the language is only spoken at home. An obvious place to start is school. Optimally the language is not just studied an hour at a time, but is actually the language of instruction.

In some areas (e.g. Canada, Catalunya) there's been a push to offer government services in minority languages. It can't hurt, but I'd argue that it doesn't help much, either: people don't interact with the government so much that this service affects people's feeling as to whether a language is needed or not.

There needs to be media in the language. The lesson of manga shouldn't be forgotten-- not just that content needs to exist, but it has to be content that people want. One Brazilian tribe had the right idea: they were given a tape recorder, they used it to record their own songs and festivals. (Thank heavens a bureaucrat wasn't around; he'd have insisted on translating government newsletters, or some such nonsense.)

Implications for parents

If you've read this far, you already know the drill: languages aren't easy even for children, and they'll learn only if they reckon they have no choice. The one certainty is that they'll grow up as fluent speakers of the street language-- the language of their peers.

If you speak a different language at home, your children will grow up speaking it; the tricky bit is when they discover that you understand the national language too. You may insist that they speak the home language at home; but results will be better the more the children hear the home language elsewhere. It's easier to retain Spanish if you live in a Hispanic neighborhood with plenty of Spanish-speaking relatives nearby. Trips to wherever the language is spoken full time and frequent interaction with monolinguals will cement the language.

It's harder if only one parent speaks the target language, especially if it's the father. My Peruvian brother-in-law, who lives in Brazil, had some idea of teaching his daughter Spanish. He knew the right approach-- speak only in Spanish at home-- but in practice found this impossible, and of course the girl is learning only Portuguese. Or consider the musician Nicolas Slonimsky, who taught his daughter Latin-- an eccentric choice, perhaps, but the baby didn't know that. Nonetheless, she eventually rebelled, demanding that he speak "the way Mommy talks".

Complicated arrangements like "speak Spanish on Mondays, Tuesdays, and Thursdays" probably won't work. As noted, children like to use a rule that a particular person speaks a particular language; it will probably only confuse them if anyone may, at seemingly unpredictable times, use any language.

How about a nanny? I've heard of this working; on the other hand, my wife used to take care of children, and none of them learned much Spanish. It's a good way for the nanny to pick up the kids' language, though.

If a language is used only at home, be aware that the children will learn only an impoverished form of it-- possibly to their surprise, if they go to college and take courses in it, expecting an easy A. A high-school friend of mine was German but grew up in the States; returning to Germany, she was embarrassed to find that her written German was terrible. Similarly, my wife, who was educated in Peru, is always distressed at the low quality of the Spanish produced by Hispanics who've grown up here. (I know a linguist whose reaction was, "But they're still native speakers." Sure, but they have a lousy command of the written standard.)