The HN headline shorterner has really run off the rails here. The actual headline is "The Mystery of People Who Speak Dozens of Languages" but HN shortens this to "The Mystery of People Who Speak Languages" which makes no sense at all.
It's really dumb in general. I'm an adult, I don't need this paternalistic protection from a headline that's trying to grab me. I also don't need constant reminders of the absurdity of our hubris that technology can solve any problem.
You make it sound as if the HN system itself arbitrarily shortens titles, which it does not. It just enforces a limit on how long titles can be. It is the user who must make the edits to fit the title into the limited space. And it was not a great editing job by the submitter.
The HN system does indeed shorten titles automatically, e.g. removing initial "How". I wish HN would stop doing this; it makes many legitimate titles nonsensical, and incidentally can remove a signal that a link originally had a clickbait-style title.
The submitter can click "Edit" to restore their original title after the autoshortener has changed it; but there is no way to tell HN "Please use exactly this title in the first place."
(I remember this submission of mine being affected by the autoshortener, for example: https://news.ycombinator.com/item?id=40114482 "How thumb indexes are cut" autoshortened to "Thumb indexes are cut".)
This is simply a UI issue. It would be totally fine if, when you went to submit a link to HN, the UI before posting said "I'm going to use this title; is that OK? [default is yes]". The only problem is that HN's UI doesn't do that confirmation step; instead it forces everyone to double-check their own post after posting. And then when someone innocently fails to do that double-check, we get these meta subthreads and the only recourse is to ask a mod to retitle the post. It's a silly aspect in a system that otherwise tends to "trust the user."
Once in a while, the HN titling system mangles a title, and people complain. What doesn't get much comment are the many hundreds of titles it improves, all the more impressive given how much effort goes into gaming titles. Cut the site some slack. You can call out suboptimal titling without building up a dudgeon.
First of all, why had nobody mentioned linguistics as a field of study by the time I had finished school? I don't think I had learned anything at all in this area until after university, and it's so incredibly rich and fascinating. I reckon the kids lose a lot by having French/German/Spanish/etc without any form of linguistics involved, where they can have a little look at the language from a different perspective.
Second, I take issue with the idea that you can't learn a language to fluency in adulthood. Like the guy in the article. I think it's really all economics:
- When you're a kid, you have no opportunity cost. Taking the time to repeat something you pronounced wrong costs you nothing, because you're just playing.
- When you're a little bit older, you want to socialize. You can socialize with your friends without speaking perfectly, and they find it awkward to correct you.
- When you're an adult, you have to work. Most jobs don't require you to speak fluently, so you don't. Jobs that require native fluency will be taken by people who had low opp cost, aka natives or early learners.
The only people who consciously try to learn a language to native level are the people in this article, who seem to be able to do it. To me they are the evidence that this would work, if only you defied your economic incentives.
The obvious parallel is coding languages. You might have dabbled in Rust as a JS or cpp guy, but once someone is paying you to write it for a living, you take the time to actually learn the idiomatic ways to do things, the libraries to use, and so forth.
It also seems obviously true that everyone can learn a bunch of languages. The old people in my family speak three languages, plus the local language of where they ended up as refugees, plus English. Most of them are not university educated, they just grew up with a bunch of languages. People who grew up at a language border tend to speak a a couple as well, and people who met a spouse from another place, eg I have a friend who learned Swedish due to moving over the bridge, and Spanish having married a Mexican.
On the computer side, writing more than one language is probably by far the norm, who doesn't write python along with something else?
Programming languages are different because you don’t have the difficulties of matching pronunciation with writing systems, aspects like literal vs. metaphorical meanings and idiomatic expressions, irregular conjugations, and the usual n:m semantic relationship between words in different languages. Literal translation between programming languages often works, and in any case is independent of whether it’s idiomatic for the language or not, whereas in natural languages it quickly becomes unintelligible.
The absolute hardest part (for me) about learning a language is listening comprehension. I can read, write, and speak French at a reasonable level, but I've never been able to understand a native speaker. It would take a lot of work and exposure to bring my mastery up to the level of my second-year written French. A large part of that is because spoken French is a hot mess compared to the written form, but I digress.
A written language without a spoken form is a lot easier to grok. Especially programming languages and their limited grammar. But I think there's more to it than that. One of the reasons I find it trivial to learn programming languages is that code is code is code, no matter which flavor of sugar you shovel on top. The language is simply an abstraction for a logical construct. Skilled programmers work on that logical construct directly, and the language is nothing more than a means to an end.
Also there's the fact that the written word is intransient. You have time to pause and search your memory or dictionary for a word. A spoken language requires extensive training to literally rewire your brain. You can't stop to ponder a word because it's gone before you can reconstruct a word from sounds. There's a very good reason that our brains have hardware to turn sounds into linguistic tokens. It's just too much processing to do at a conscious level at any reasonable speed.
As a counterpoint, my sister learned German in college, going on to get a Master's and becoming fluent --- rather envy her that and often wonder what would have happened if I had chosen (or been offered the choice of) a different language when making the selection when I enlisted.
Similarly, the French Foreign Legion seems to be quite successful at teaching French.
> However they differ, the hyperpolyglots whom I met all winced at the question “How many languages do you speak?” As Rojas-Berscia explained it, the issue is partly semantic: What does the verb “to speak” mean?
When hearing about hyperpolyglots I always have some doubt. Some people believe they know a language when they have a vocabulary of a few dozen words. That might be enough to order food at a restaurant but I don't think is what most people would consider "speaking a language". How many of the 22 languages Rojas-Berscia "has command of" can he speak at a highly-proficient level like CEFR C2? I'm guessing not many.
The question of linguistic distance is also rarely touched upon in these articles despite being of critical importance. It's far easier to learn Italian when you already speak French, Spanish, and Portuguese. A quantitative analysis of hyperpolyglots ought to have some sort of "linguistic range" measure to capture the distance between the different languages you speak proficiently.
This is what I always wonder in the back of my head as well. Take passive vocabulary size - it should be fairly easy to show people who claim to speak a given language a random selection of words in various frequency bands, and ask them either what they mean or what other words out of a list are closest to it semantically, to estimate their actual vocabulary size. If you don't think showing isolated words is an adequate metric of someone's true gloss skill, you can show words embedded in sentences. But we never even hear about these kinds of estimates in these stories, let alone anything more in depth.
C2 is about much more then language itself. It is a lot more about presentation skills, writing specific kind of formal text and such. It is truly an unreasonable expectation for someone saying they "speak a language".
As a person who grew up speaking his mother's tongue, learned the language of a country where she had been born as the child of a worker under the National Mobilization Law, and who learned English from his father and Armed Forces Radio and TV, ultimately forgetting the other two languages at his father's behest, failed miserably at his one year and one day of high school French, and when sent to the Presidio of Monterey when I enlisted to re-learn my mother tongue only had a small smattering come back and who never managed to regain conversational fluency, it is quite mysterious to me how some folks learn languages so readily.
Computer languages seemed a match for this facility (despite being the token graphic designer in the comparative programming languages course I was the only one who managed to do all of the Lisp problems successfully), but these days I find myself dragging blocks around to rough out designs using: https://www.blockscad3d.com/editor/ rather than directly programming in OpenSCAD (or even the new Python variant: https://pythonscad.org/ ) and wonder if it is a consequence of my choosing to express myself through drawing rather than by words when I was young.
I thought about it and realized that i have a lot of friends and colleagues who speak more than 3 languages. Many even speak 5 or more.
One reason is that many have two mother tongues, e.g. Russian and Ukrainian or Russian and Lithuania, or Spanish and Catalan or swiss German and Italian.
Another reason is that they have emigrated to a country where they had to learn the language.
Also, in most European, it's normal to learn two or three languages in secondary school, beside your native language, and at least one other. Combinations like English, Latin, and Spanish; English and French; or English, Spanish and German are common as secondary languages. By learning two or three languages in school, and already speaking one or two native languages, you know between two to five languages by the time you finish school.
Basically, you learn English at school everywhere. often even before that, because tv and other media are not translated into the local language. That's for example common in the Nederlands or Scandinavian Countries.
Now, if you like languages and have already learned a Romance language at a young age, for example, it is much easier to learn other Romance languages later. This means that if you already speak Spanish or Portuguese, it is easier to learn Italian.
This is something that really only seems like a mystery when viewed from an explicitly academic context. Speaking many languages is the norm for many people, and speaking many, many, languages is a very refined ability. When you think of language as "purely" mental and private then it seems "mysterious", but once you rid yourself of the mentalist (Cartesian in philosophical terms) confusion and focus on language as public behavior, then it's not mysterious at all.
I don't see how thinking of language as public behaviour explains anything about how some people are able to learn many languages very easily and some struggle to learn an additional one.
I don't think anyone is able to learn many languages very easily. If you want to have meaningful discussions in a language you need a vocabulary on the order of 5-10k words depending on the subject and how much exposure you have, which is a cool year of study for even the fastest learners and more akin to 2-3 for us mere mortals.
Most cases of people speaking 10+ languages lean heavily on speaking many closely related languages in a family, which most people already cope with pretty well.
Most people that know many languages, know many languages because they're trying to do things with many languages, and they grow up in that situation: they have to speak many languages. You need to get things done, so you communicate in the way that allows you to do that. They're not thinking about grammatical rules, in fact many people that learn many languages almost never learn grammar very well at all, what they learn is how to ask for a bathroom, where an address is, what's such and such called, or how do you say..., or how are you today?, how much is this?, that is okay, and various responses, ect., things that allow them to do things with people who speak different languages. Often the languages are from similar families and the peoples are from similar regions. I've met many people from the indian sub continent, for example, who can speak 4+ languages, or people from the European alps who can speak 3 or 4, quite well (which does not mean grammatically accurate in a school sense). The people that learn wildly different languages are devoting extensive amounts of time to it and are outliers. They obscure the reality that speaking many languages is a part of different communities interacting and trying to get stuff done together. It's not a private mental exercise, it's a part of public life for many.
>They're not thinking about grammatical rules, in fact many people that learn many languages almost never learn grammar very well at all, what they learn is how to ask for a bathroom, where an address is, what's such and such called, or how do you say..., or how are you today?, how much is this?, that is okay, and various responses, ect., things that allow them to do things with people who speak different languages.
Learning phrases is not learning a language. Again, you're doing nothing to explain the very real phenomenon of individual differences in second language acquisition.
My mother tongue is English, but I learned French and German for my work. My French colleagues say I "speak French" and my German colleagues say I "kinda speak German."
Would you say they're wrong? When, according to you, has someone learned a language? You don't get a medal at some point, you learn a language when you can do things with it.
Read a transcript of the words used in a typical TV show. Then read a book written for adults (not the pulp fiction ones).
I sometimes discover that I mispronounce words. The problem is I'd only read them, and assumed they were phonetic.
> language requires much more than just vocabulary
The vocabulary is all you need to "get along". I did not claim that makes you a native speaker, or enables subtlety in word order. I've traveled in foreign countries, and make an effort to learn a few words. It's surprising how far you can get with just a handful. (P.S. making an effort to communicate in the local language wins lots of friend points! I highly recommend it.)
Pick up one of those phrase books for various languages. They aren't that deep, and they work.
I do not know why this is downvoted. If you know tv 2000 words and can use sentences like they do in TV, you are already in pretty good situation. You wont be able to discuss complex topics, but you will be able to get by and communicate more of what you need.
Of course there will be a room for improvement ... but you will be able to get along in foreign language speaking environment and bootstrap from there.
The HN headline shorterner has really run off the rails here. The actual headline is "The Mystery of People Who Speak Dozens of Languages" but HN shortens this to "The Mystery of People Who Speak Languages" which makes no sense at all.
Makes no sense and is not needed. The original is short enough.
Seemed clear enough. The use of plural “languages” already made it a mystery to me!
It's really dumb in general. I'm an adult, I don't need this paternalistic protection from a headline that's trying to grab me. I also don't need constant reminders of the absurdity of our hubris that technology can solve any problem.
> I also don't need constant reminders of the absurdity of our hubris that technology can solve any problem.
Yeah, we get enough of those from the submissions.
You make it sound as if the HN system itself arbitrarily shortens titles, which it does not. It just enforces a limit on how long titles can be. It is the user who must make the edits to fit the title into the limited space. And it was not a great editing job by the submitter.
The HN system does indeed shorten titles automatically, e.g. removing initial "How". I wish HN would stop doing this; it makes many legitimate titles nonsensical, and incidentally can remove a signal that a link originally had a clickbait-style title.
The submitter can click "Edit" to restore their original title after the autoshortener has changed it; but there is no way to tell HN "Please use exactly this title in the first place."
(I remember this submission of mine being affected by the autoshortener, for example: https://news.ycombinator.com/item?id=40114482 "How thumb indexes are cut" autoshortened to "Thumb indexes are cut".)
Here's another recent one: https://news.ycombinator.com/item?id=44312099 "Lawyers are mad about SALT" quietly turned into "Lawyers are mad about salt."
This is simply a UI issue. It would be totally fine if, when you went to submit a link to HN, the UI before posting said "I'm going to use this title; is that OK? [default is yes]". The only problem is that HN's UI doesn't do that confirmation step; instead it forces everyone to double-check their own post after posting. And then when someone innocently fails to do that double-check, we get these meta subthreads and the only recourse is to ask a mod to retitle the post. It's a silly aspect in a system that otherwise tends to "trust the user."
Once in a while, the HN titling system mangles a title, and people complain. What doesn't get much comment are the many hundreds of titles it improves, all the more impressive given how much effort goes into gaming titles. Cut the site some slack. You can call out suboptimal titling without building up a dudgeon.
The original title is short enough for HN. I'm assuming "dozens of" was removed because it sounds like a listicle.
Random thoughts after skimming this...
First of all, why had nobody mentioned linguistics as a field of study by the time I had finished school? I don't think I had learned anything at all in this area until after university, and it's so incredibly rich and fascinating. I reckon the kids lose a lot by having French/German/Spanish/etc without any form of linguistics involved, where they can have a little look at the language from a different perspective.
Second, I take issue with the idea that you can't learn a language to fluency in adulthood. Like the guy in the article. I think it's really all economics:
- When you're a kid, you have no opportunity cost. Taking the time to repeat something you pronounced wrong costs you nothing, because you're just playing.
- When you're a little bit older, you want to socialize. You can socialize with your friends without speaking perfectly, and they find it awkward to correct you.
- When you're an adult, you have to work. Most jobs don't require you to speak fluently, so you don't. Jobs that require native fluency will be taken by people who had low opp cost, aka natives or early learners.
The only people who consciously try to learn a language to native level are the people in this article, who seem to be able to do it. To me they are the evidence that this would work, if only you defied your economic incentives.
The obvious parallel is coding languages. You might have dabbled in Rust as a JS or cpp guy, but once someone is paying you to write it for a living, you take the time to actually learn the idiomatic ways to do things, the libraries to use, and so forth.
It also seems obviously true that everyone can learn a bunch of languages. The old people in my family speak three languages, plus the local language of where they ended up as refugees, plus English. Most of them are not university educated, they just grew up with a bunch of languages. People who grew up at a language border tend to speak a a couple as well, and people who met a spouse from another place, eg I have a friend who learned Swedish due to moving over the bridge, and Spanish having married a Mexican.
On the computer side, writing more than one language is probably by far the norm, who doesn't write python along with something else?
Programming languages are different because you don’t have the difficulties of matching pronunciation with writing systems, aspects like literal vs. metaphorical meanings and idiomatic expressions, irregular conjugations, and the usual n:m semantic relationship between words in different languages. Literal translation between programming languages often works, and in any case is independent of whether it’s idiomatic for the language or not, whereas in natural languages it quickly becomes unintelligible.
That's an interesting point.
The absolute hardest part (for me) about learning a language is listening comprehension. I can read, write, and speak French at a reasonable level, but I've never been able to understand a native speaker. It would take a lot of work and exposure to bring my mastery up to the level of my second-year written French. A large part of that is because spoken French is a hot mess compared to the written form, but I digress.
A written language without a spoken form is a lot easier to grok. Especially programming languages and their limited grammar. But I think there's more to it than that. One of the reasons I find it trivial to learn programming languages is that code is code is code, no matter which flavor of sugar you shovel on top. The language is simply an abstraction for a logical construct. Skilled programmers work on that logical construct directly, and the language is nothing more than a means to an end.
Also there's the fact that the written word is intransient. You have time to pause and search your memory or dictionary for a word. A spoken language requires extensive training to literally rewire your brain. You can't stop to ponder a word because it's gone before you can reconstruct a word from sounds. There's a very good reason that our brains have hardware to turn sounds into linguistic tokens. It's just too much processing to do at a conscious level at any reasonable speed.
As a counterpoint, my sister learned German in college, going on to get a Master's and becoming fluent --- rather envy her that and often wonder what would have happened if I had chosen (or been offered the choice of) a different language when making the selection when I enlisted.
Similarly, the French Foreign Legion seems to be quite successful at teaching French.
Learning multiple programming languages is practically trivial compared to human ones.
Programming languages have a “vocabulary” of what, 20-50 keywords? And modern languages have extremely similar “grammar”…
Mastering 500 (even irregular) verbs and 500 nouns of a human language is fairly useless for interacting with other humans beyond very limited scopes.
> However they differ, the hyperpolyglots whom I met all winced at the question “How many languages do you speak?” As Rojas-Berscia explained it, the issue is partly semantic: What does the verb “to speak” mean?
When hearing about hyperpolyglots I always have some doubt. Some people believe they know a language when they have a vocabulary of a few dozen words. That might be enough to order food at a restaurant but I don't think is what most people would consider "speaking a language". How many of the 22 languages Rojas-Berscia "has command of" can he speak at a highly-proficient level like CEFR C2? I'm guessing not many.
The question of linguistic distance is also rarely touched upon in these articles despite being of critical importance. It's far easier to learn Italian when you already speak French, Spanish, and Portuguese. A quantitative analysis of hyperpolyglots ought to have some sort of "linguistic range" measure to capture the distance between the different languages you speak proficiently.
This is what I always wonder in the back of my head as well. Take passive vocabulary size - it should be fairly easy to show people who claim to speak a given language a random selection of words in various frequency bands, and ask them either what they mean or what other words out of a list are closest to it semantically, to estimate their actual vocabulary size. If you don't think showing isolated words is an adequate metric of someone's true gloss skill, you can show words embedded in sentences. But we never even hear about these kinds of estimates in these stories, let alone anything more in depth.
C2 is about much more then language itself. It is a lot more about presentation skills, writing specific kind of formal text and such. It is truly an unreasonable expectation for someone saying they "speak a language".
B1 then. The point is that there should be some sort of objective measure other than simply "Hey how many languages do you speak?"
(2018)
And the title should be “…who speak many languages.”
Speaking one language may be something of a mystery but to most of us it’s at least a familiar state.
https://archive.ph/RJ6pK
As a person who grew up speaking his mother's tongue, learned the language of a country where she had been born as the child of a worker under the National Mobilization Law, and who learned English from his father and Armed Forces Radio and TV, ultimately forgetting the other two languages at his father's behest, failed miserably at his one year and one day of high school French, and when sent to the Presidio of Monterey when I enlisted to re-learn my mother tongue only had a small smattering come back and who never managed to regain conversational fluency, it is quite mysterious to me how some folks learn languages so readily.
Computer languages seemed a match for this facility (despite being the token graphic designer in the comparative programming languages course I was the only one who managed to do all of the Lisp problems successfully), but these days I find myself dragging blocks around to rough out designs using: https://www.blockscad3d.com/editor/ rather than directly programming in OpenSCAD (or even the new Python variant: https://pythonscad.org/ ) and wonder if it is a consequence of my choosing to express myself through drawing rather than by words when I was young.
I thought about it and realized that i have a lot of friends and colleagues who speak more than 3 languages. Many even speak 5 or more.
One reason is that many have two mother tongues, e.g. Russian and Ukrainian or Russian and Lithuania, or Spanish and Catalan or swiss German and Italian.
Another reason is that they have emigrated to a country where they had to learn the language.
Also, in most European, it's normal to learn two or three languages in secondary school, beside your native language, and at least one other. Combinations like English, Latin, and Spanish; English and French; or English, Spanish and German are common as secondary languages. By learning two or three languages in school, and already speaking one or two native languages, you know between two to five languages by the time you finish school.
Basically, you learn English at school everywhere. often even before that, because tv and other media are not translated into the local language. That's for example common in the Nederlands or Scandinavian Countries.
Now, if you like languages and have already learned a Romance language at a young age, for example, it is much easier to learn other Romance languages later. This means that if you already speak Spanish or Portuguese, it is easier to learn Italian.
This is something that really only seems like a mystery when viewed from an explicitly academic context. Speaking many languages is the norm for many people, and speaking many, many, languages is a very refined ability. When you think of language as "purely" mental and private then it seems "mysterious", but once you rid yourself of the mentalist (Cartesian in philosophical terms) confusion and focus on language as public behavior, then it's not mysterious at all.
I don't see how thinking of language as public behaviour explains anything about how some people are able to learn many languages very easily and some struggle to learn an additional one.
I don't think anyone is able to learn many languages very easily. If you want to have meaningful discussions in a language you need a vocabulary on the order of 5-10k words depending on the subject and how much exposure you have, which is a cool year of study for even the fastest learners and more akin to 2-3 for us mere mortals.
Most cases of people speaking 10+ languages lean heavily on speaking many closely related languages in a family, which most people already cope with pretty well.
Most people that know many languages, know many languages because they're trying to do things with many languages, and they grow up in that situation: they have to speak many languages. You need to get things done, so you communicate in the way that allows you to do that. They're not thinking about grammatical rules, in fact many people that learn many languages almost never learn grammar very well at all, what they learn is how to ask for a bathroom, where an address is, what's such and such called, or how do you say..., or how are you today?, how much is this?, that is okay, and various responses, ect., things that allow them to do things with people who speak different languages. Often the languages are from similar families and the peoples are from similar regions. I've met many people from the indian sub continent, for example, who can speak 4+ languages, or people from the European alps who can speak 3 or 4, quite well (which does not mean grammatically accurate in a school sense). The people that learn wildly different languages are devoting extensive amounts of time to it and are outliers. They obscure the reality that speaking many languages is a part of different communities interacting and trying to get stuff done together. It's not a private mental exercise, it's a part of public life for many.
>They're not thinking about grammatical rules, in fact many people that learn many languages almost never learn grammar very well at all, what they learn is how to ask for a bathroom, where an address is, what's such and such called, or how do you say..., or how are you today?, how much is this?, that is okay, and various responses, ect., things that allow them to do things with people who speak different languages.
Learning phrases is not learning a language. Again, you're doing nothing to explain the very real phenomenon of individual differences in second language acquisition.
My mother tongue is English, but I learned French and German for my work. My French colleagues say I "speak French" and my German colleagues say I "kinda speak German."
Would you say they're wrong? When, according to you, has someone learned a language? You don't get a medal at some point, you learn a language when you can do things with it.
Something I read long ago:
English has 1 million words
College graduates have a 30,000 word vocabulary
High school graduates have a 10,000 word vocabulary
TV vocabulary is 2,000 words
What this means to me, is that I can get along in another language by learning only 2,000 words.
Even assuming that’s true, speaking a language requires much more than just vocabulary.
> Even assuming that’s true
Read a transcript of the words used in a typical TV show. Then read a book written for adults (not the pulp fiction ones).
I sometimes discover that I mispronounce words. The problem is I'd only read them, and assumed they were phonetic.
> language requires much more than just vocabulary
The vocabulary is all you need to "get along". I did not claim that makes you a native speaker, or enables subtlety in word order. I've traveled in foreign countries, and make an effort to learn a few words. It's surprising how far you can get with just a handful. (P.S. making an effort to communicate in the local language wins lots of friend points! I highly recommend it.)
Pick up one of those phrase books for various languages. They aren't that deep, and they work.
I do not know why this is downvoted. If you know tv 2000 words and can use sentences like they do in TV, you are already in pretty good situation. You wont be able to discuss complex topics, but you will be able to get by and communicate more of what you need.
Of course there will be a room for improvement ... but you will be able to get along in foreign language speaking environment and bootstrap from there.