Critical periods for language | Hacker News Reader

Critical periods for language

astralcodexten.substack.com

24 pointsfeross21 comments

Regarding the drop in learning speed after school, I have a pet theory that primary and secondary education are less about teaching people specific details of educational topics, but rather giving them a wide range of meaningful vocabulary in their native language to talk about those topics and others. As soon as you enter the workforce or even a college education, your range of meaningful topics to discuss dramatically narrows as you focus on a specialty or profession.

I wonder if perhaps one reason that young children might be able to acquire language quicker is that they're less worried about making mistakes and sounding foolish. I know it's definitely a barrier I have to overcome when trying to speak foreign languages in the real world.

Kids are lavished with positive attention when they make mistakes, because it's cute.

With one exception, I've never believe the theory that kids learn better,* and I have data, though on a small n. My mum's side of the family is heavily polyglot of necessessity; her parents moved to a new country (where the languages came from different language families) and my grandparents were completely fluent in at least one local language. I also married someone who was not an English speaker and her english is more complex than that of most people here in the US.

The critical factor is use. If you learn a language only in school you don't really use it. Languages I spoke daily as a kid: mostly gone, except English, my daily drive. If you're embedded in another culture you're going to need the language to get by and you'll continue learning.

* For me the exception is accent. Kids learn what phonemes matter and that does seem sticky. But in terms of grammar, vocabulary, slang and semantics, I have not observed those who learnt as kids being any better than those who learnt as adults and used it.

One of my kids learned a foreign language by watching cartoons in that language on youtube. We had NEVER spoken that language at home (but I can speak it). They didn't speak that language in kindergarten (he was 5 years old at that time and we didn't live in a country where that language was spoken). He just watched the videos and after a few months we noticed that he was laughing "at the right places". We asked him why he was laughing. He said "Because the [main character] just said [something specific about a horse]". I thought "okay, maybe he had watched that episode already in our language and he remembered what they said", so I asked him "Can you say [something in our language] in [the foreign language]?". He did. It was correct, pronunciation and grammar.

Do you think an adult could do the same?

Sure, I don't see why not if they were willing to put in the same amount of time your kid did. That is basically how you acquire language when you move into a culture. The value of classes, IMHO, is marginal, especially before you have some gross concept of the structure of the language.

I have never seen an adult being exposed just one hour per day to a foreign language over three month, and then suddenly producing grammatically correct meaningful sentences. And note that we are talking here about passive exposition. No teacher, no corrections, no feedback, no interactions, no carefully crafted pedagogical input, not even slowed down audio.

If you know such an adult, please tell them to write a blog about it. I am curious. I have met many adults in multilanguage environments and even some who, as a kind of scientific experiment, specifically tried that approach, without much success though (after some weeks, they could recognize some words and fixed expressions).

I don't say that an adult cannot learn a foreign language by immersion. Actually, there are probably hundreds of millions of such people in the world (take any uneducated worker who is forced to work in a foreign country to help his family survive at home). But the conditions are very different (they get feedback and interactions) and the quality of the result, too. The speed and quality with which kids just absorb a new language like sponges is unmatched.

Yes, there are books based on this method, for example Lingua Latina per se Illustrata

Not until the comment section will you find reflection on time you can actually dedicate to language acquisition. Even on unconscious level, adults are generally already far more overloaded with many other topics to deal with.

From my recollection of my developmental psych courses, the studies on this subject generally control for "time available" by including cohorts of adults in full-time immersion programs, or even adults who move permanently to a new country. The results have always consistently shown that even with dedicated study and full time immersion, adults will never reach the same level of linguistic fluency as would a child within the critical period (generally < 11 years old) because they're incapable of acquiring the phonemes of a new language to the same degree as a child, due to fundamental limits of neuroplasticity in the regions of the brain responsible for language acquisition.

So yes, obviously if you don't have time as an adult to learn a language, you won't learn it as well as if you can dedicate the rest of your life to it. But the point of the results of "critical period" studies is that even if you dedicate the rest of your life to it, you'll still be at a disadvantage compared to a child, when it comes to attaining perfect phonetic and linguistic fluency.

Does that suggest that you could arrange a kind of course that all children got, which taught the union of all phonemes of all common languages, without teaching much of the languages themselves? And thereby future-proof the child with the bits you can't learn as an adult, in case they wanted to learn any particular language later?

Probably yes, and I think this is what a lot of rich people do with their kids, at least in terms of exposing them to vastly differing languages like Mandarin and English.

I do wonder if there's some limit on the number of phonemes you can learn, and how many hours are required to "solidify" a phoneme. But I definitely think the idea of a "unity course" is a good one.

Personally I had a French "class" in elementary school, where I barely learned anything - I certainly can't speak French - but I can pronounce French words better than I can pronounce any Spanish words (despite three years of Spanish class in college). Those French classes were basically just learning to count to 100 and singing songs like "alouette."

I'm always skeptical of these poorer language learning results.

It took me 3 years of about 2 hours a day in high school to learn approximately 6000 words in English, my native language. This was in addition to my schooling where I was writing essays and book reports pretty much weekly. Probably 10+ hours per week for history and english. Plus all the interactions in my native language.

That's a LOT of language practice. That's almost 1800 hours on vocabulary alone. Add on 1500 hours for essays. That's approaching 3200 hours which strongly exceeds the amount of time required to gain general proficiency in a Category V language (2200 hours). That's almost 10+ hours per week on top of school hours themselves.

How many people ever put in that amount of effort into a second language? How many people have the time to put in that amount of effort into a second language?

I dunno, if you're an adult who moves from India to America, you're not going to be speaking Hindi much (unless you speak it at home, to be fair). You'll be conducting your life entirely in English. And if you move to the States for university then you'll be doing academics in English too. You'll easily achieve thousands of hours of practice in this way.

While most people don't have time to put that many hours into studying a second language outside of a country where it's the native tongue, many immigrants have no choice but to immerse themselves in a new language just by the default behavior of living their life in a new country that doesn't speak their first language.

Yes, such immigrants will always be a few thousand hours behind people who started in America, but surely there's some plateau where the improvements become marginal. I don't think the critical period advantage is about the early volume of hours so much as the age period when they're spent - i.e. if you compare someone who moves at 20 vs. 40, the results will be the same, but both will be worse than someone who moved at 8.

I've re-read your comment multiple times and still don't quite get it. If English is your native language, why did you spend 1800 hours learning English vocabulary when you were in high school? Did you feel that your native level was not sufficient, or were you just interested in advanced English as an academic interest?

As a native English speaker, I'm not sure how many English words I know, nor how I would even start counting! However, I am currently learning German, and seem to have found a comfortable pace of learning about twenty words each day, studying for half an hour each day. This doesn't include general exposure to the words through other German practice. If it is possible to keep that up long-term, that means that 6000 German words would take just under a year for me to learn in the ideal case (probably a bit longer, as so-called 'strong' words have multiple irregular forms that are effectively different words).

> If English is your native language, why did you spend 1800 hours learning English vocabulary when you were in high school?

Mostly SAT prep. It took my verbal score from 550 to 780 (out of 800).

However, extra vocabulary is also one of the fastest ways to improve your writing. My writing got a lot more concise the more vocabulary I commanded. Essays that used to take 20 hours+ and maybe 6-8 pages were now 4-5 hours and 2-3 pages.

And, once you do this, you find out how very few words we actually use on a daily basis.

“The difference between the almost right word and the right word is really a large matter. ’tis the difference between the lightning bug and the lightning.” -- Mark Twain

I'm intrigued as to what you mean by 'incapable of acquiring the phonemes of a new language'. If that is the only limit on adult language acquisition, then it seems like it would follow as a native English speaker, I should be able to learn Japanese to a higher level than a native speaker of Japanese would be able to learn English. That is because Japanese only has a few sounds that English doesn't, whereas English has a lot more sounds than Japanese.

Something about that doesn't feel right, so I hope you will be able to explain if I am missing something. However, were it to be true, it sounds like the best way of 'futureproofing' children's early language education so that they can learn languages later in life is to bring them up to speak, at least at a basic level, a set of languages that covers a large proportion if not all of the distinct phonemes of all languages.

I suspect you're right that an adult will be able to achieve "nearer" phonetic fluency in a language that shares more phonemes with their first language than they would in one that has little phonetic overlap with it. So a Spanish speaker will have an easier time becoming fluent in Italian than would an English speaker, etc. (I'm surprised by your assertion that Japanese has large phoneme overlap with English, but I don't know much about it - it just seems surprising to me.)

I'd be curious to see a visualization of the overlapping phonemes between languages, like a large Venn diagram kinda thing (or maybe just a graph). I think this data is probably available and a visualization wouldn't be too difficult to create - maybe it already exists but I couldn't find a comprehensive one with a brief search.

I love the idea of a visualization of overlapping language phonetics! something that showed both the hierarchical nesting of languages and the actual phonemes themselves.

Besides tsu and dzu, every phoneme in Japanese is accounted for in the Spanish language. I believe that the Spanish phonemes are a nearly perfect subset of English phonemes.

> If an English-speaking adult was placed in a monolingual Spanish family, in a part of Spain with no English speakers, after a few years they might find they’d learned Spanish “easily” too.

That is not so. That person would have an accent, struggle to say as well as understand things and make weird grammar mistakes that a four-year-old native speaker wouldn't.

Oh, and how about being immersed in a household speaking a language that is not even remotely related to English?

A child which speaks nothing and is learning Spanish natively is not comparable to someone who already knows English, which is in the same Indo-European family and has tons of congnate words with Spanish.

Language learning is correlated with socialization. We learn by communication with other people. As people get older, it seems they have a harder time making new friends and knowing new people. I wonder if this explains part of it. Perhaps the average adult learner, even when fully immersed in a foreign language environment, doesn't socialize with other unfamiliar adults as readily as a child would do if left to play with other children. Less socialization = less language practice = slow learning.

Reply on news.ycombinator.com