native level

[jkhsu]

You really think that the 17 year old will do significantly better?

I actually think the 17 year old might not even be able to read the article in his mother language!

But I do agree with you that, pronunciation aside, given the exact same environment and study conditions, the adult can do just as well on the other aspects of the language.

[Silent]

Based on my observations of Chinese / Chinese Americans I have met in the USA: I've met many ethnic Chinese who immigrated to the USA around age 5 or so and have near perfect English accents. At the same time, I haven't met any ethnic Chinese who immigrated to the USA during college years or later who have near perfect English accents

Sorry, but I think just accent is an extremely limited evaluation of language skills. Language skills are a combination of pronunciation, reading, vocabulary, grammar, actual understanding including all kinds of nuance, and probably a bunch more. If it's only accent you're after, yes, children are at a big advantage over adults. If you consider overall skills I have my doubts or children are still at an advantage. For short term (couple of years) study I think adults are at an advantage. Over time children may get the advantage as general language skills will largely be decided by personaility/talent/effort. After all, even many natives have poor language skills. So over time vocabulary, grammar etc will average out at a (near) native level and pronunciation will tip the balance towards the children.

[jkhsu]

Sorry, but I think just accent is an extremely limited evaluation of language skills

I tend to agree as well if we're just talking about accents alone. English is spoken with many different accents depending on the region you are from. I've even said before in other posts that there are plenty of people from say Singapore and India who, I think, have a much better command of English than many Americans in the USA.

The part that I question is whether learning the language as a child has any more benefits related to speaking beyond accents? Does age affect how you think in that language? Does age affect how you pick up usage of grammar? These questions are more unclear to me.

[daofeishi]

This is one of those things, like the origin of religion or the state of the current economy, where everyone has a pet hypothesis. Even though opinions are interesting, I'd like to see less of them and more hard data. Many aspects of language learning are not that hard to quantify and get reliable data on. We just have to agree on what we are discussing

1. Do adult learners actually reach the same level as young learners (regardless of reason)?
   
2. Do adult learners do better or worse at aspect X of language learning under identical conditions?
   
3. Can adult learners do as well or better at aspect X of language learning than children if they are given enough time/resources?
   

I think the evidence points to the negative in cases 1 and 2, and to "maybe" in case 3. Of course we are talking about averages here. Children born with defective FOXP2 genes will we worse learners than adults, and a talented learner of Emil Krebs-like abilities might do better than the average child, but that has no bearing on the issue. We also have to be careful not to fall into a false innate trait vs. acquired trait dichotomy. There are many traits in humans and other animals that have important components of both, e.g. where learning can only happen in a critical phase or where there are ideal phases to acquire a given trait.

A very illustrative example if you don't buy that language in humans could be such a trait, is the acquisition of stellar navigation in indigo buntings. Steve Emlen et. al. found that learning the orientation of the night sky could only happen in nestlings, and those birds that were deprived of stimulus at that stage were forever unable to navigate (look up Emlen's paper, it's a fascinating read.) Similar critical conditions are found in humans as well. People who are born with congenital cataracts and are deprived of early visual stimulus but who are later in life given an operation never learn to utilize their vision the way ordinary adults do. Several important abilities usually are not developed, such as the ability to tell distinct geometric objects apart. The point of these examples is that it is not warranted to be skeptical of the claim that adults are worse language learners than children because one does not believe in traits that are solidified at an early age. There are neurological developmental processes that are mostly developed in critical stages, and a priori we cannot exclude that language or important components of language acquisition are an instance of that.

That adults as a practical matter on average do not become as proficient in a language they learn late in life should not be a point of contention, and Steven Pinker references many studies on this in "The Language Instinct," e.g. SAT/GRE scores among adult learners with 20+ years of experience and aptitude tests controlled for age of onset and language exposure. He also writes that

More systematic evidence comes from the psychologist Elissa Newport and her colleagues. They tested Korean- and Chinese-born students and faculty at the University of Illinois who had spent at least ten years in the United States. The immigrants were given a list of 276 simple English sentences, half of them containing some grammatical error like The farmer bought two pig or The little boy is speak to a policeman. (The errors were errors with respect to the spoken vernacular, not ''proper'' written prose.) The immigrants who came to the United States between the ages of three and seven performed identically to American-born students. Those who arrived between the ages of eight and fifteen did increasingly worse the later they arrived, and those who arrived between seventeen and thirty-nine did the worst of all, and showed huge variability unrelated to their age of arrival.

There are many aspects of language learning that adults simply do not perform as well on as children do. The first aspect, although probably far from the most important one in practical terms, is that accent acquisition is provably hard after approximately age 10. Not only is acquisition a critical trait, but the ability to learn new phonemes has a huge developmental component to it as well. Elizabeth Spelke found that the ability to tell phonemes whose distinction is not found in your own mother tongue apart diminishes enormously with age. The example that I was presented was the ability to tell the Hindi consonants da and Da apart. British, American, French and Israeli toddlers all had a 99% success rate in telling them apart without previous exposure. In adults over the age of 18, even with practice, the rate dropped to less than 20%. Even in adult learners with considerable exposure to the language, recognition rates were nowhere near that of the toddlers. This was referred to in Pinker's book as well:

English-learning infants under the age of six months distinguish phonemes used in Czech, Hindi, and Inslekampx (a Native American language), but English-speaking adults cannot, even with five hundred trials of training or a year of university coursework. Adult ears can tell the sounds apart, though, when the consonants are stripped from the syllables and presented alone as chirpy sounds; they just cannot tell them apart as phonemes. (p. 264, The Language Instinct)

As a matter of phonetics, the evidence shows that the age of onset does count. What about grammar? Referring to the results mentioned in the aptitude tests above, adult learners do perform worse than those who learned the language early on, even given the same level of exposure. That this is likely to have to do with the criticality of grammar development in children can be found in the study of pidgins, i.e. languages that are constructed in communities where people from many different ethnic backgrounds come together and have to find some way of communicating. The adults who first construct the proto-language, the generation 0 pidgin, always construct a language with little grammatical regularity and a lot of syntactical confusion. What is really fascinating is that when their children learn the language and we get the "first generation native speakers," their grammar becomes standardized and conforms to universal rules of grammar found in other languages. (The pidgin is said to be creolized.) This is fairly strong evidence that grammar development is an innate ability in children that tends to atrophy in adults.

The perhaps most studied example is from Nicaragua, when the first schools for the deaf were opened in the 70s. When the first cohort came together, there was no standardized sign language in Nicaragua. All that existed was a collection of "homesigns" constructed by the children themselves to communicate with their families and friends. Now that they came together for the first time, children of all ages, they had to find a way of communicating, and a language developed organically among them. What was found was that in the language that developed, the older children were not able to proceed past the somewhat constrained and ungrammatical system om homesign, whereas the younger children in the group developed a highly grammatical and structured sign language with much greater nuance. (Ref. Kegl & Lopez, "The Deaf Community of Nicaragua and their sign language(s)") A similar account is found in The Language Instinct, where Pinker mentions a child born to deaf parents who had acquired imperfect ASL late in life:

Astoundingly, though Simon saw no ASL but his parents' defective version, his own signing was far better ASL than theirs. He understood sentences with moved topic phrases without difficulty, and when he had to describe complex videotaped events, he used the ASL verb inflections almost perfectly, even in sentences requiring two of them in particular orders. Simon must somehow have shut out his parents' ungrammatical "noise." He must have latched on to the inflections that his parents used inconsistently, and reinterpreted them as mandatory. And he must have seen the logic that was implicit, though never realized, in his parents' use of two kinds of verb inflection, and reinvented the ASL system of superimposing both of them onto a single verb in a specific order. Simon's superiority to his parents is an example of creolization by a single living child.

In the Nicaraguan example, we have younger and older children in the same cohort that spent the same amount of time on learning the language under the same conditions and with the same motivation. Still, the younger ones were the only ones who showed any ability to construct a coherent grammar. I find this very compelling evidence that grammatical understanding is something innate in children, and something that adults on average will be much worse at.

Since this post is getting fairly long, I will have to refer back to my previous post on vocabulary acquisition. Children have a remarkable retention rate for words and language constructs, which adults don't have, and this shows that in three huge areas of language learning, adults do perform worse than children, even under the best possible circumstances.

I know there are many people who find it demotivating to think about, but the people who frequent forums like this one are not in general among the worst language learners out there. Some will probably want to cite cases of their own and other people's successes in defense. But we have to be careful about generalizing from our own experience and from anecdotes. Only carefully crafted studies can tell us what the situation is like, and I think Pinker's case for the existence of a critical language instinct is well made.

I am fairly sure that case 1 and 2 have to be answered in the negative. Adult learners on average do worse, and have less of a knack for languages than children, even if you put them in an environment where all they did was learning the language as optimally as possible.

Putting fingers in our ears singing "na na na" will not make that fact go away. What we instead have to do is to optimize our language learning and strive towards being evidence of case #3 above.

Edit: I tried to find a good citation for my claim about vision in cataracts patients, and discovered this. The claim that vision is a critical ability seems to be debated, so I'll have to leave that to the experts. I hope my point still remains clear, that it could be possible that language learning has a critical period similar to navigation in buntings.

[jono1001]

I agree I am sceptic about the fluency. It reminds me of the claims of the Irish Polyglot blog Fluent in Three Months. I am always sceptical of these type of claims.

[rezaf]

The main question is what kind of adult we are talking about. As I said in another post we might lose some of our impulsive language learning abilities but in return we get more analytical abilities and it all depends on how hard we work on them. I don't know anything about those older kids in Nicaragua or their education background but probably the reason many people on this forum would not accept that children are faster in learning languages is that many of the members have reasonably high education background and many know quite a few languages already. Who do you think can learn faster; a seven year-old kid or a 30 year-old who can speak 5 languages and has got one or two university degrees?

[anonymoose]

Regardless of whether children can learn languages faster or not, I think one has to be cautious when comparing language abilities to those of an educated native speaker. What do we actually mean by "educated native speaker"? Does an educated native speaker necessarily have better language than a not-so-well educated native speaker? Many so-called educated native English speakers in the UK still make mistakes in English, whilst many speakers who have not experienced higher education speak very well. Is one's language level measured by how technically (grammatically) correct it is? Or by the range of vocabulary used? Or by the specialist topics that can be discussed in the said language (which then involves the knowledge of the topic as well as the language)?

I think that it may be possible (albeit extremely difficult) to reach a "native-like" level in a foreign language when we compare the language as used in normal everyday situations. In other words, it may be difficult to distinguish the native speaker from the non-native speaker in the setting of an informal chat. But I refuse to believe that a learner could start from scratch and, within two years, truely achieve the same level as an educated native speaker when we consider all the specialist and infrequently used words and expressions that a native speaker picks up through 10+ years of academic study. I don't believe it's possible, simply because even if one has phenomenol language learning skills and a memory like a computer, two years is simply not long enough to cover all the material covered by a native speaker in 10+ years.

There may be some foreign learners of English who could speak informally and be indistinguishable from me, but I doubt they would be familiar with such words as stomata, corrie, thalidomide, iambic and so on just to name a few random examples of infrequently used words that I learnt during my school education.

[Silent]

Many so-called educated native English speakers in the UK still make mistakes in English, whilst many speakers who have not experienced higher education speak very well.

I think it's fairly obvious that a well educated person has more scope of making language errors. A well educated person makes more complex sentences and conveys more complex ideas, as a simple consequence it's more likely to make a mistake. There's little that can go wrong if you say "you are wrong", if you say "you did well on x but there is some scope for improvement on the y aspects of the analyses and therefore the conclusion you have come to is of doubtful quality consequently further research is necessary before we can take any decisions on the subject" there are more opportunities to make mistakes.

two years is simply not long enough to cover all the material covered by a native speaker in 10+ years.

Pretty much the argumentation I used before. But then, what do you consider well educated? Do you really need to known the words that are learned by a native over an x-year period while the words have become rare? If a student has exactly the same size vocabulary as a native of the same age I think the student has a bigger useful vocabulary and has reached a higher vocabulary level. The native will be loaded with useless 'extinct' words.

[realmayo]

Interesting post daofeishi, to highlight that it's grammar as much as accent that young children do better.

Maybe I should have put it like this: small children don't learn a language, they learn language.

As adult learners, being discouraged by this is daft: it's like a 30-year old being discouraged that certain private body parts are not getting any bigger the same way they did during puberty.

In fact, on the basis that the evolutionary process has caused our species to develop in a way (which was) best suited to our survival, there was no need for people to learn languages which they weren't exposed to after the age of, say, 10. Or at least, no real benefit. Which doesn't mean there isn't benefit now, and doesn't mean that we can't used our intelligence to find a way of learning new languages which is different from children. The same way that the above-mentioned 30-year old who wants bigger private parts can go to a doctor. Or humans who want to fly can get in an aeroplane. But, obviously, we have to find a "smart" way to do this; an adult trying to recreate the language environment of a child would, by this reasoning, fail.

[renzhe]

daofeishi, that is an excellent post with well-picked references. Thanks for that, I gave you a green plus.

Since I'm not a neurophysiologist, I do not want to start arguing each of your points as I'd be getting out of my field and that's when internet arguments get silly. But we might be talking past each other, so I'll try to briefly clarify my point.

There is no hard, physical evidence that language learning is a skill we lose with age. We have such evidence for many aspects of cognition, but not for language learning.

There is hard evidence that that people born blind actually reprogram large parts of their visual cortex for other tasks, such as audio-based localisation. We have CAT scans and electrode readings. They can never learn geometric shapes again because they have never learned any shapes to begin with, and that part of their brain is used for a different purpose now. But people who were not born blind do, in fact, learn new shapes as long as they are alive. You never lose this ability.

There is hard evidence explaining the difficulty of obtaining good accents later in life. There is muscle memory making the creation of new sounds more difficult. There is also the tuning of neuron populations to certain stimuli which occurs over decades, and makes it increasingy difficult to detect new sounds -- at the expense of far better discrimination of known sounds.

Nobody has ever provided hard evidence that there is a change in the brain which causes adults to stop being able to acquire grammar and vocabulary. Electrodes, CAT scans, brain imaging... This is understandable, considering that language is perhaps THE highest cognitive ability of all, and these are very difficult to measure with electrodes. These are also something you can't measure in monkeys or cats. Still, until I see such evidence, I'll continue being sceptical.

In the meantime, there are dozens of perfectly good environmental reasons why adult immigrants do worse on average: amount of exposure, discrimination, huge amounts of work and life-related stress, language interference, much higher expectations from peers, higher level of self-criticism, etc. I mean, Chinese and Korean graduate students in the US are notorious for not interacting with the locals. That they do worse than pre-school children with 8-10 hours per day of forced interaction with other children does not prove a neurophysiological inability to learn.

[realmayo]

There is hard evidence that that people born blind actually reprogram large parts of their visual cortex for other tasks

There is also hard evidence that people who don't learn any language as a child but start learning as an adult use the opposite side of their brain that children use to learn language: why's this? It suggests that the way adults learn language is different from how children learn language!

Nobody has ever provided hard evidence that there is a change in the brain which causes adults to stop being able to acquire grammar and vocabulary

... ... does not prove a neurophysiological inability to learn.

It seems like you're setting up something of a straw man here. I don't think anyone says that impediments suddenly come down on adults making it impossible to learn languages!

Just that there is a period in a human's development when the brain develops the ability to use language (not a language, but language), and that's when learning languages happens really "easily", without conscious effort. Like: sex organs start developing at puberty. No one would argue that these things are at their most potent soon after developing, I mean, compare the "activity" down there of an 18-year old with an 80-year old!

This idea that development of brains is not nature but all nurture is surely wrong. Yes, brains adapt all the time. But part of that is programmed: it's how humans develop. For instance, the brains of teenagers are different from those of children and adults. The parts of the brain that focuses on the self is more active. That's why teenagers change personality and become, for a while, basically more selfish. (According to something I read -- could be rubbish of course!)

The brain is just a part of the body. We accept that the body develops in broadly predictable patterns as people get older: growth spurts, puberty, etc etc. I accept it might be a bit scary to think that the brain does the same thing -- that some of the changes to our way of thinking are not because of external influences, but are "preprogrammed", as it were. But while it may be scary, it makes perfect sense if you accept that the brain is part of the body, rather than some magical "spirit" etc etc.

[jkhsu]

Is near perfect Mandarin pronunciation off the table for most learners here since most started learning Mandarin as adults? Don't think there are ＜10 year old kids here right?

[Silent]

Define 'near perfect pronunciation'. I mean a while back you pretty much equaled non-native accent with language skills. If native accent is the measure yes near perfect becomes extremely hard. If 'near perfect' is as close or closer to a standardised pronunciation than a significant part of the population it gets far more attainable as Chinese dialects vary widely.

[realmayo]

jkhsu: for a start, the fact that it is easier for a child to learn a language than an adult wouldn't mean that an adult can't learn a language as well as, or better than, a child -- it just means it would be more difficult. That's all anyone's saying isn't it? That it's easier for children than adults.

[renzhe]

There is also hard evidence that people who don't learn any language as a child but start learning as an adult use the opposite side of their brain that children use to learn language

Everybody learns a language as a child. Some people learn two, some three, some four.

This is a completely different situation to somebody who has never learned to see, or never learned stellar navigation, etc. and is missing a part of the brain development as a result. We all learned a language, so the corresponding areas of the brain are well developed by adulthood.

I completely agree that a child who grows up into adulthood having never communicated with anyone using language will likely be unable to learn any language in their adulthood. I also agree that adults learn differently, but this does not mean that their learning is inferior.

This idea that development of brains is not nature but all nurture is surely wrong.

That's why nobody is suggesting such a thing. This seems like a false dichotomy to me.

If adults lose the ability to acquire language -- and we are talking about huge changes -- as they age, then there should be neurophysical evidence for it.

Is near perfect Mandarin pronunciation off the table for most learners here since most started learning Mandarin as adults?

Absolutely not.

There are a few foreigners with near perfect Mandaring pronunciation (Da Shan, for example), and as soon as learning Mandarin becomes widespread, there will be many more.

[anonymoose]

I also agree that adults learn differently, but this does not mean that their learning is inferior.

If one accepts that adults learn differently to children, then surely one must also accept that their learning ability is different to children. I mean, although learning ability is difficult to quantify, it must in principle still have a value, and thus excepting the infinitesimally small possibility that adults' and children's learning ability is equal, then one must necessarily be inferior to the other. Of course, that doesn't resolve the argument of which age group is better, but logically, you cannot claim that children learn differently to adults yet maintain that both have an equal learning capacity.

[renzhe]

Sure, but nobody is interested in infinitessimal differences in ability to learn languages.

What is claimed is that children learn languages considerably more easily and considerably better than adults. Which is why I used the word "significantly" in a previous post.

There are most likely all kinds of differences. But I am not convinced that adults are inherently (=physiologically) bad at learning languages and that children have it "easy". An average child spends 10x more time learning a language than the average adult learner. If they are 10x as effective, I suggest that this might be the first place to look, not brain changes which remain hypothetical to this day

[jkhsu]

Define 'near perfect pronunciation'.

When I say near perfect pronunciation, what I mean is that the learner's pronunciation is the same as native speakers from a particular location (the location where the learner spent most of his/her time learning). In other words, the learner does not have a "foreign" accent. For example, someone from Shanghai can have a Shanghai accent when they talk to people in Beijing but in Shanghai, they would be normal. A better way to describe this is if I talk to this person over the phone, I can't tell that they are an English speaker. I might think they they are from Beijing or Shanghai but not from another country. That's near perfect pronunciation to me.

it just means it would be more difficult. That's all anyone's saying isn't it?

For the most part I agree. However, with pronunciation, I have my doubts. Sure there are always outliers and people who are gifted at speaking. But an average adult language learner will have difficulty achieving a level such that they are indistinguishable over the phone from a native speaker.

There are a few foreigners with near perfect Mandarin pronunciation (Da Shan, for example), and as soon as learning Mandarin becomes widespread, there will be many more.

I am not at the level to judge Da Shan's pronunciation but I have heard native speakers say that even his pronunciation is distinguishable from a native Chinese speaker.

Seriously though, I know there are many advanced learners here but how many of you are indistinguishable over the phone when compared to a native Mandarin speaker? And I don't mean just saying some stock phrases or highly rehearsed conversation.

I wouldn't be surprised if there are some at this level just because there are some seriously gifted learners here but this is really the exception and not the norm. However, if you grew up in China as a kid or were born in China, it would be more of the norm for you to be indistinguishable from a Chinese native over the phone.

Perhaps most of us are better qualified to judge non-native English speakers speaking English in our own countries. How often have you encountered talking to a Chinese immigrant who started learning English as an adult over the phone and have no idea they were not a native English speaker from your country? I've mentioned in another post that I haven't. That's why I find it hard to believe it's anything "normal" that the other way around can be done by adult learners of Chinese, even with time and effort.

[Silent]

. An average child spends 10x more time learning a language than the average adult learner.

Though not 10 times It may and I suspect it to be true that children spend more time learning language in immersion environments, but on average this is of course just not true. I started learning English in secondary school. Nowadays English lessons are started at primary school. I don't think there is a significant difference in time per week spend on learning. I think most children just spend the language lessons learning the language, make their homework and that's it. Probably the average adult learning a language is spending more time. An adult will learn a language with a real purpose and may be more motivated to spend a couple of hours extra beside what's compulsory for school. I definitely spend more time on my chinese studies then I spend for my compulsory highschool lessons French and German together.

[realmayo]

"immersion" is a red herring! If a child has parents from different countries and is also looked after by someone from a third country, and they all spend time talking to the child, it will speak three languages. This isn't immersion in one language. If such a child is spending this putative 10x the time an adult is studying language, and the child is awake for 15 hours per day, supposedly "studying" non-stop, then one divides 15 by 10 by 3 to get, what half an hour? So find me a 30-year old who studies for 30 minutes a day, and compare them after 10 years with a 10-year old......

Or: give 100 random 30-year olds plenty of resources to learn another language. See how many of them are fluent after 10 years.

Find 100 random 0-year olds. See how many of them are fluent after 10 years.

Also: I just happened to hear this on TV. If chimp doesn't learn these skills before the age of 4, it never will. Sound familiar?

Back on that other thing:

A: There is also hard evidence that people who don't learn any language as a child but start learning as an adult use the opposite side of their brain that children use to learn language

B: Everybody learns a language as a child. Some people learn two, some three, some four.

No -- I (A) was referring to brain scans of someone who had not learned any language as a child.

Finally:

but this does not mean that their learning is inferior

Look, we're all guessing aren't we? But for me, the hypothesis that humans have a period in their life where they learn language, as well as a period in their life where they learn motor skills, as well as a period in their life where their sex organs develop, say, seems perfectly logical. And it has the virtue of explaining what is pretty self-evident to most people: that children learn languages quicker than adults. My own experience: arriving at a Chinese university at the same time as some under-10s who spoke no Chinese: ask any of the security guards, the children learned quicker than the adult students, despite often having less exposure to Chinese.