Aug 29 2011

The Bilingual Brain

A new study sheds further light on how we learn language, showing the power of early exposure to language. Language is something that humans are very good at, which is a way of saying that we have cortex dedicated to and specialized for language.

The dominant hemisphere (dominant by definition, the left hemisphere in most people) contains several structures specialized for language. Wernicke’s area is the brain’s dictionary – it translates words into concepts and concepts into words. When you are trying to think of a specific word for something – that process is taking place in Wernicke’s area. There is also specialized cortex that processes auditory information, translating sounds into words, and feeding that information to Wernicke’s area. In the frontal lobe there is Broca’s area that converts words into speech – essentially this is specialized motor cortex that allows for the exquisite control of the muscles of speech necessary to produce the subtle sounds of speech. Wernicke’s area and Broca’s area are connected by a cable called the arcuate fasciculus.

It has also been known for a long time that the language cortex develops when we are very young, beginning in infancy and then pretty much locking in place by the time we are four years old. This is the “window” of development for language. For children who are raised without exposure to language by age four, they will never acquire normal fluency. If you normally developed a primary language, and then learn a second language after age four, the second language will use more brain resources – it won’t have the same language cortex encoding that the primary language does.

This window of development also applies to phonemes – the individual sounds that make up words. We learn a finite number of phonemes by age four, and then we are pretty much stuck with that set of language sounds. Children as young as 8-10 months old start to learn how to distinguish sounds that are part of speech from other sounds, and the speech sounds (phonemes) become encoded in the language cortex. After about four years old no new phonemes can be learned, and every speech sound we hear from that point forward will be slotted into an existing phoneme.

This is where the new study comes in. The researchers found that babies 8-10 months old were already developing the ability to distinguish sounds important to language – like the difference between an “r” and an “l” in English. However, they also found that babies not exposed to a language that makes this distinction, for example like a child raised in a Japanese household, will start to lose the ability to make this distinction at the same age. So by 10 month old children are already losing their ability to learn new languages, or at least the sounds that make up languages.

In addition they found that children raised in a bilingual household had a greater facility for learning language and the window of learning new sounds was extended. The researchers hypothesize that in a bilingual environment children learn that there is more than one word for everything, and they also greatly stimulate their language cortex by the task of switching between the two languages.

This study extends what was already known about the advantages of learning multiple languages by age four. This study, in fact, indicates that if you want to maximize your child’s language ability, they should be exposed to a second language from infancy.

The study looked at children raised in a bilingual household. It is unclear if the advantage extend to simply playing a second language for the children to hear, but that is not being used interactively with another person. Perhaps a video would be better than pure audio, but still that should be studied specifically before any claims are made for such products. Children may need direct human interaction to make the connection with language.

This study does add to existing evidence, and suggests that perhaps parents should be encouraged to expose children to more than one language. Language instruction might also be added to preschool and daycare programs. By the time children reach Kindergarten, it’s already too late.

28 responses so far

28 Responses to “The Bilingual Brain”

  1. Tim Schmelteron 29 Aug 2011 at 8:56 am

    Very interesting post. I’ve heard language development discussed before in the context of neuroplasticity, but two things about this post struck me.

    First, “If you normally developed a primary language, and then learn a second language after age four, the second language will use more brain resources – it won’t have the same language cortex encoding that the primary language does.” Does that mean that children raised from infancy in a bilingual environment have both languages encoded in the “primary” area? And does their increased future proficiency in learning new languages suggest they encode those new languages differently?

    Second, “After about four years old no new phonemes can be learned, and every speech sound we hear from that point forward will be slotted into an existing phoneme.” Is anybody aware of audio samples that play two different phonemes that, when heard by native speakers of different languages, sound the same? I’ve wondered if there were phonemes that I, as a native English speaker, cannot detect in a manner similar to the canonical example of a Japanese speaker’s inability to distinguish “l” vs. “r”. Unfortunately, my Google searches only turned up articles and demonstrations of speech recognition software.

  2. nybgruson 29 Aug 2011 at 9:10 am

    I learned 3 languages right off the bat and went to school to learn my 4th in 1st grade. Interestingly enough, I was able to learn the phonemes of my 4th language (French) quite well – to the point, in fact, where I can speak completely without accent. I suppose that learning the first 3 essentially at the same time extended that window for phonemic recognition.

    Now, later in life, I find that I can pick up and piece together many other languages very easily… as long as they are in the same family as the others I speak (Slavic, Germanic, or Romantic). After just 3 weeks in Bavaria I was understanding and speaking traditional Bavarian better than some of the North Germans I was hanging out with (sadly, it went just as quickly as it came). But Asian languages are still a mystery to me and even with significant practice I was unable to hammer down the phonemes there – I always have an accent when I muddle through the few phrases of Chinese I know.

    Sadly, I’ve rarely gotten the chance to practice anything but English for the past few years, so everything else is becoming rusty. But it does come back rather quickly when I become immersed.

    Most importantly, however, I still retain my ability to order a beer in 10 different languages (and say cheers as well). We all have to keep our priorities straight, after all.

  3. ChrisHon 29 Aug 2011 at 11:35 am

    My first language was Spanish. I was born in the Panama Canal Zone and my parents employed a maid (very common in those days, the late 1950s). My parents spoke English to me, while the maid spoke to me in Spanish. Then my father was stationed back in the USA when I was three.

    I could speak Spanish as well as any three year old and some English. Moving back to the States was apparently upsetting to me, my parents tell I stopped speaking (elective mutism) for several weeks (or months depending on parent).

    In a few years I essentially forgot Spanish. But when I was eleven my father was stationed in Venezuela where I went to an American school with a daily Spanish class (level dependent on skill, I was a beginner).

    I was able to do well pronouncing the words, though the grammar is another thing. Between two years in Venezuela and two years of high school Spanish (the minimum required, in defiance to my father) I can speak some Spanish, and it is mostly trying to pull up vocabulary and remembering the tenses that trips me up.

    And when I speak my “native” language, English, I keep getting asked where my accent came from. There seems to be some holdover from the early Spanish immersion. It is strong enough to really muck up my attempt to learn French, it seems that I really want to pronounce all the letters and do it phonetically!

  4. HHCon 29 Aug 2011 at 11:51 am

    A bit of history, in 20th century Chicago, parents were told to speak only English in the home to prepare children for grammar school classes. After the World Wars, German-speaking families
    in Iowa had their language removed from the curriculum in schools.
    Political correctness extended to the home and was enforced at school.

  5. kijibajion 29 Aug 2011 at 12:29 pm

    Some of Pat Kuhl’s earlier research suggests that social interaction is a key component for infants in second/foreign language learning: audio and audiovisual material was not enough to maintain infants’ ability to perceive non-native phonemic contrasts (

  6. Steven Novellaon 29 Aug 2011 at 12:45 pm

    Cool – thanks for the link

  7. Mark Ericksonon 29 Aug 2011 at 12:56 pm

    “By the time children reach Kindergarten, it’s already too late.”

    It’s too late for innate cognition – easy brain work. You can still be fluent in another language (with an accent, perhaps), it just takes lots of work: time, effort and brain resources. Is that right?

  8. tmac57on 29 Aug 2011 at 12:58 pm

    I guess that I had always assumed that the problem that the Japanese had with pronouncing the ‘r’ sound was due to not learning the mouth and tongue movements necessary to make the sound,rather than not being able to hear the difference. There are many sound in the Spanish language that I can recognize,but have trouble reproducing when I speak. Trilling R’s for example.

  9. ccbowerson 29 Aug 2011 at 1:19 pm

    I’m not sure that it is all so simple. My wife did not speak English until she moved to the US at age 6 (up until that point she spoke only Japanese). By the time she was a teenager she says she was not nearly as comfortable speaking Japanese as she was with English despite her parents very limited English speaking ability.

    Only after working in a Japanese airline did she acquire proficiency in her first language. We are now raising our children bilingually right off the bat (I speak only English), and even now despite being fluent in Japanese my wife still feels more comfortable with English.

    My oldest daughter is ~5 and is so far remains bilingual in both English and Japanese. Only now is she beginnning to show some preference to English due to school, however she remains approx equally proficient in both

  10. ccbowerson 29 Aug 2011 at 1:31 pm

    “The researchers hypothesize that in a bilingual environment children learn that there is more than one word for everything, and they also greatly stimulate their language cortex by the task of switching between the two languages.”

    Witnessing this “switching” in a toddler was quite amazing. As soon as she could speak, my daughter knew when to switch with pretty good accuracy.

    “By the time children reach Kindergarten, it’s already too late.”

    I agree that earlier the better, but it seems discouraging to say that we need to start before school. Its not really feasible for most people. My daughter has taken Spanish in preschool, but there is only so much they can teach in that setting. I know many people who have learned language proficiently after this time period, and I think it would be great to start in kindergarten (as opposed to commonly ~8th grade)

  11. elmer mccurdyon 30 Aug 2011 at 1:25 am

    But the cutoff age is generally considered to be around 9, not 4, based on observing behavior (I don’t know how scientifically), rather than examining the brain. It’s certainly consistent with my experience teaching kids to speak English.

  12. Shelleyon 30 Aug 2011 at 9:14 pm

    I’m wondering whether early learning of a language itself is necessary or if the simple sounds will do. For example, does being raised by a grandparent who speak english, but with a heavy accent make a difference? Lots of early sound is a kind if imitative babble and not real words. Learning to trill r sounds or learning a guttural g might be part of listening and imitating sounds and not language itself.

  13. LivingWithMormonson 30 Aug 2011 at 9:18 pm

    I’d be curious to understand this a little better.

    I was born in Brazil and moved to the States at age 15. I’m 36 now, but very quickly became proficient in the language. After about a year or so I had a Northeast accent since that’s where I learned English. After moving to Florida and recently Utah, I have no discernible accent.

    I dream and think in English and consider myself as fluent in it as Portuguese (sometimes even more).

    “If you normally developed a primary language, and then learn a second language after age four, the second language will use more brain resources – it won’t have the same language cortex encoding that the primary language does”
    “By the time children reach Kindergarten, it’s already too late.”

    It’s too late for what?

  14. borealyson 31 Aug 2011 at 12:25 am

    I’ve wondered if there were phonemes that I, as a native English speaker, cannot detect in a manner similar to the canonical example of a Japanese speaker’s inability to distinguish “l” vs. “r”.

    There’s an online IPA chart with audio samples here:

    One that’s actually pretty easy to hear if you listen for it is the aspirated vs unaspirated voiceless stop consonants (for instance, the /t/ in “top” as opposed to the /t/ in “stop” — in “top,” there’s a little puff of air following the /t/. In “stop” there isn’t.) In English we treat those two variants as being the same, to the point where we don’t even notice the difference. In some languages (including Mandarin Chinese), they are perceived as being as different as /t/ and /d/.

    Interestingly, even though we don’t consciously notice the difference, a person who doesn’t use the proper pattern of aspirated vs. unaspirated sounds will sound noticeably odd to most native English speakers — like they have a mild accent. (In French, consonants are almost never aspirated the way they are in English. My learned-it-well-past-preschool approximation of the accent improved quite a bit once I figured that one out.) I’m currently providing speech therapy to a little girl who, among other articulation problems, can’t aspirate her stop consonants.

    Aaaand I realize I just wrote two paragraphs on aspirated consonants.

    One distinction I never did manage to learn to perceive? The various back fricative sounds in Arabic (for instance, the h, the hooked h, and the crossed h in the IPA chart). My Arabic-speaking classmates found it self-evident that the professor was playing three different sounds, but to me, they all just sounded like /h/. Still do — if I listen to them one after the other, I can hear that they’re different, but damned if I can tell which is which without peeking.

  15. nybgruson 31 Aug 2011 at 3:20 am

    @borealys: Interesting read indeed. I had a roommate once who was into linguistics and we chatted casually about such things. I found it very interested, though never enough to supplant my medical interests so I just never went down that rabbit hole too much. One thing we did figure out is that our other roommate (from Iran) could never pronounce the “th” sound in English. We figured out (empirically – mind you were were sophomores in undergrad at the time) that to make “th” you must put your tongue past the front of your teeth. In Farsi, there are no phonemes that require you to do that manipulation with your tongue. Once we figured that out, he could pronounce the sound perfectly… until he didn’t concentrate on it and slipped back into his accent.

    @shelley: I reckon it does make a difference. I know of people personally (and have seen documentaries tangentially related to the topic) that cannot speak any language other than American English but have a heavy accent still. Back in my home of Southern California, I met many children Mexican immigrants who did not speak Spanish to save their lives (heck, I knew more often!) but still had a heavy Mexican accent when they spoke.

    I reckon that just exposing them to the phoneme set is insufficient since, as my Iranian roommate demonstrated for us, it isn’t just knowing the phoneme that is important but also how to use it and remembering to do so. It would be interesting to see though if, say, a French child was only taught French but exposed to English phonemes from birth would they then be better able to learn English later in life with less (or no) accent?

  16. techczechon 31 Aug 2011 at 2:01 pm

    I’m a bit shocked by the complete lack of any critical balance in the blog post or even a hint of skepticism. Nothing mentioned about the critical period has been uncontested. Plus making conclusions based on a press release is highly questionable.

    Kuhl’s research is extremely narrow and does not reflect anything we actually know about language learning (first and second). I’ve written a detailed critique about her conclusions here:

    In short, there’s definitely no such thing as a narrowing window on phoneme distinction that’s anywhere close 10 months. Children at the age of 7 (and even adults as LivingWithMormons demonstrates) are still perfectly able to acquire the phonetic apparatus of a new language or several languages. Kuhl’s research shows at best their ability to participate in highly contrived laboratory experiments.

  17. ccbowerson 31 Aug 2011 at 8:44 pm

    “It’s too late for what?”

    Thats what I was wondering. I believe that he means that it is too late to learn a another language that your brain treats in the same way as the first language. There seem to be examples in which this appears to not be practically significant, and thats when I referenced my wife who did not learn English until age six, but it is now her (slightly) stronger language. I imagine this senario is common for those who move to another country as a youngish child.

    “In Farsi, there are no phonemes that require you to do that manipulation with your tongue.”

    I find that odd, because a long time friend of mine’s family is from Iran and I grew up hearing Farsi a lot. I donot recall the “th” sound in Farsi, but I clearly remember this mother referring to “clothes” in English with the soft form of “th,” almost in an over exaggerated way. In the NE US that word is pronounced with hardly any “th” sound at all, often pronounced similarly to “close,” with a slight hard “th” sound.

  18. ccbowerson 31 Aug 2011 at 8:55 pm

    “I’ve wondered if there were phonemes that I, as a native English speaker, cannot detect in a manner similar to the canonical example of a Japanese speaker’s inability to distinguish “l” vs. “r”.”

    I did not realize that there was a difficulty with distinguishing the sounds, I assumed it was due to the distinct sounds not existing in the Japanese language. (Or perhaps these two concepts are part of the same problem). In japanese there is one sound that is not quite r, not l and harder sounding than the english form of either. People tend to think of foreign words/sounds in the sounds of their first language, and I wonder if the difficulty stems simply from this. I’ve always go the impression that they can hear the difference, but it gets confusing because the difference is not something they are used to distinguishing

  19. nybgruson 31 Aug 2011 at 10:39 pm


    I could be wrong or it could have been his particular dialect? But we sat around one evening and had him make as many Farsi phonemes as he could recall and it was his assessment that such a phoneme did not exist in his language. He would always hit his “th” with a hard “/t/” sound.

  20. borealyson 31 Aug 2011 at 11:37 pm


    I think, in general, yes, most people can hear the difference between different allophones of the same phoneme — but only if they’re paying attention. The difference isn’t meaningful, so most of the time, we just don’t notice it.

    After all, there are countless factors affecting the acoustics of an individual sound when we talk. What we perceive as one phoneme is actually an uncountable set of acoustically distinct sounds influenced by everything from the speaker’s health to the other phonemes in the word. The brain has to make sense of the set if it ever hopes to learn language, so, very early in life, it starts sorting the noise into groups based both on similar acoustic features, and the specific distribution of sounds in the language(s) it is exposed to. Each language draws the lines in different places. (There have been some really neat cross-linguistic studies on vowels that look at this.) The brain is, by necessity, wired not to notice acoustic differences it doesn’t need to notice, and it decides what it needs to notice based on experience during the first couple of years of life.

    This is also why, as you say, people tend to perceive foreign phonemes in terms of the phonemes of their own language. I’m always amazed at the number of French-English bilinguals who don’t realize that the French R and the English R are not only different, but radically different. A French R is closer to an English H than to an English R, but because they are used in ways that are similar with respect to *meaning* (in cognate words such as treasure-tresor), and because they are written with the same letter, people don’t always notice the difference unless they stop to think about it — even when they regularly speak both languages, with both Rs. But really … they don’t need to notice, any more than they need to notice the difference in aspiration between English and French T sounds.

  21. nybgruson 01 Sep 2011 at 1:19 am


    That makes sense to me. It explains why I could understand my father through his very, very thick Russian accent with no issue at all and my friends would all be lost trying to understand him. It was odd, sitting in the same room I could understand my native english speaking friends just as clearly as my father, and while I could detect he had an accent it didn’t seem even remotely as bad as it did to my friends.

  22. ccbowerson 01 Sep 2011 at 8:04 am

    “I could be wrong or it could have been his particular dialect?”

    Well I wasn’t thinking that he was wrong really. It could be that she was going out of her way to make that sound because it wasn’t part of Farsi. I have noticed that in other people as well, when a sound is not part of their first language. For example, on more than one occasion heard a Japanese airline announcer who overpronounced her “r” sounds.

    “I think, in general, yes, most people can hear the difference between different allophones of the same phoneme — but only if they’re paying attention.”

    That is the impression that I get from most senarios that I’ve encountered with others. It is hard to tell sometimes because hearing a difference is not the same as being able to replicate that difference when one speaks (although I assume that the former is required for the latter).

  23. Kultakutrion 01 Sep 2011 at 9:03 am

    I wonder, how does language acquisition work later on? Because, I hear learning languages is difficult but I never experienced any difficulty, to a continuous awe of passers-by, who never cease to be fascinated that I learned Finnish just for the heck of it, without any major problems. I feel it’s like playing with a kitten – takes time but it sort of happens and if you force it, the joy of feeling the brain working and building new synapses (metaphorically) goes away.

  24. borealyson 01 Sep 2011 at 9:14 pm


    I’ve heard of people who can learn a new language with native-like proficiency even as an adult — my undergraduate phonetics professor claimed to have had one as a PhD student. It’s a rare skill, especially when it comes to getting the accent right. If you have it, it *is* impressive.

    It’s not impossible or even uncommon to become highly proficient in a second language once the so-called critical period has passed (my French is a good example), but it’s extremely rare for someone to reach the point where a native speaker can’t tell that you’re not one of them. Any number of things can give you away — grammatical quirks, idioms, word choice, prosody — but most often, at least in my experience, it’s the phonology. Those darn vowels betray me every time.

  25. D.P.on 03 Sep 2011 at 11:52 am

    It is not the first research that shows the ability to discriminate allophones decreases after 1 y.o. or so. However, it is a completely wrong to to say that by the time children reach Kindergarten, it’s already too late. Too late for what?

    Most kids who were immersed in a second language environment at age 6 or earlier will develop native like pronunciation even without additional instructions or training. Also, it is possible for older children to achieve that but may require special training.

    Another thing that I would like to notice is that the “window” of opportunity for learning language has nothing to do with phonetics skills. For instance, deaf children may not be able to learn any oral language, but if they are exposed to any sign language, he or she will develop language skills, which allows not only to communicate using this sign language, but also learn some other languages later. On the other hand, those who were not exposed any language before age 4 are very unlikely to become truly proficient in any language.

    “If you normally developed a primary language, and then learn a second language after age four, the second language will use more brain resources – it won’t have the same language cortex encoding that the primary language does.”

    What language requires more resources depends more on one’s proficiency than age of acquisition. Here is a result of one study:

    “These findings suggest that, at least for pairs of L1 and L2 languages that are fairly close, attained proficiency is more important than age of acquisition as a determinant of the cortical representation of L2.”

  26. Kultakutrion 06 Sep 2011 at 3:32 am

    @borealys: I guess it can be done, after all, it happens to me all the time. It’s probably related to the fact that I read everything that has letters and listen to things and at a point it just all clicks together.

    Oddly enough, I always had problems at school. I didn’t have any learning difficulty but the way my brain works is that I get pieces of knowledge and put them together. Giving me an axiom to learn by heart and exercises to understand how said rule works in reality, which is how school works, is extremely difficult to absorb. Now, there’s a thing I was explained by my mother who has Ph. D. in theory of teaching or some such so I don’t have quotes and footnotes to provide but it makes sense: Most people’s brains work analytically – they’re given instructions and rules and learn how these apply to singular cases. The small minority, around 5 %, works the other way, from observations and, so to say, case series, to generalizations, which happens to yours truly, and the school system is not exactly prepared for people who think in the opposite metaphorical direction. This might be an outdated idea or something that has no support in other research but it makes sense to me, I have problems to explain stuff to people because they don’t get my line of reasoning – in the sense that they, to me, seem to pay no attention to things around them until shown, or just don’t seem to think.

    @ D. P.: well, the article discussed was apparently about hearing things and subsequent minutiae of language acquisition so as it happens, it rules out deaf children.
    Anyway, I hear that even the deaf use dialects, slang etc. in sign language which is outside the textbook scope, and I wouldn’t wonder if similar processes of acquisition of communication mode (think cultural mannerisms of all sorts) worked the same way as verbal language.

  27. IHon 19 Feb 2012 at 11:48 pm

    This post confuses me. I know that the “window” of language development closes after a certain age in your early youth, but I don’t think phoneme recognition goes away at about 4. I’m a second-generation immigrant, and I didn’t start learning English until about age 4, when I started pre-K. Until age 4, I only knew and spoke Spanish, but it was gradually replaced with English at school. I don’t have trouble distinguishing or reproducing any of the 43 phonemes found in American English, but for some reason I transferred some English rules into Spanish, e.g., pronouncing the U’s in words like lengua, huevo, as W’s, as in language (LANG-gwij). To me they sounded like W’s whenever I would hear a native speaker pronounce these and other words (maybe because I write down everything I hear in my head, hoping for a cognate), but I found out just recently that they are actually just U’s, as in the oo sound in soup. I was very embarrassed. I know I have an accent in both languages, I’ve been told, but English remains my strongest because of my education. The point is that I was able to learn and distinguish new sounds that were not in my primary language, such as the ‘th’ sound and schwa. Now the question is, when do we lose this ability? I know I’ve lost if for sure, but don’t exactly know when because I never bothered learning a third language.

  28. IHon 19 Feb 2012 at 11:58 pm

    I forgot to include that in my experience listening and trying to understand other romance languages, such as: Portuguese, Galician, Italian and Catalan has not been hard, but making the sounds is. I don’t 100 % understand everything in a conversation when I listen to the radio or songs, but I get the gist of it, especially if I can read a transcript. I tried once reading something in Portuguese to a native speaker, and I was corrected, when I seriously thought I had nailed the sounds, so my only problem is speaking.

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.