Hacker News Reader

Comment by stanrivers | Hacker News Reader

This is interesting; I studied Korean in college / lived there for a bit. Randomly, met a Chinese woman there, who actually lived in the States, and we got married after we both moved back to the States.

If I could do it over again, I would learn Chinese before Korean or Japanese. My Chinese friends who learned Korean learned it way faster than any western learner.

Japanese students learned Korean even quicker than the Chinese students - in addition to using Chinese words, Japanese and Korean have the same grammar structure ( Subject Object Verb ) whereas Chinese and English are different (Subject Verb Object)

Learning Chinese is like learning Latin before learning a Romance language.

The added benefit is that Chinese isn’t a dead language.

Both Japanese and Korean are distinct from Chinese and beautiful languages (I really do love Korean, such a pretty, logical language), but China has been an influence in the region for so long they can’t have help absorbing and using pieces of the Chinese languages.

Fun random fact: a lot of Korean words are built on Chinese words; however, some of those words were taken from Chinese thousands of years ago.

The Chinese pronunciations have changed. Some researchers have been known to use Korean pronunciations of words to help triangulate to what ancient Chinese characters would sound like (in combination with poems where you know what rhymes to expect etc)!

View on HN

I totally agree with this analogy, though maybe to add some nuance (or pedantry), I'd claim that to start learning Chinese, you need to pick a modern dialect (probably Mandarin), which is like choosing Spanish over Portuguese or Catalan. Then, learning Japanese or Korean (or Vietnamese) is like learning a nearby but non-Romance language that was heavily influenced by Latin or Romance languages, like English. Knowing Spanish will let you appreciate the similarities with English, but you'd still need to delve into historical Spanish (or French) to really see the connections.

Chinese is a pretty old and continuous language, so at the advanced level, learners could delve into Classical Chinese, which is truly analogous to Latin, and was also a dominant lingua franca in written form across East Asia. Pieces of Classical Chinese are 'alive' in everyday use, kind of like how Latin is 'alive' in the Romance languages through general and obvious similarities, or specific vocabulary, (e.g. "quid pro quo", "ad hominem"). But Classical Chinese is radically different from any dialect of today's vernacular Chinese.

In any case, the similarities and overlap definitely make learning between CJK(V) easier!

100% agree with this post, which I tried to get at saying that Korean and Japanese are distinct. Apologies for not making that clear. Your post is great to clarify this all more specifically.

The cultural side of things definitely requires a lot of time. And that’s fun.

What also takes time is learned how ideas are shared. Not just grammar, but which version of the multiple correct grammar options that is more common. How are paragraphs constructed? What is usually shared first in an idea or argument versus the end?

You can have someone with completely correct grammar and pronunciation that sounds odd to a native speak because the cultural / idea organization is not “native”.

Go back to middle Chinese, of which, for example, the Shanghai based Wu dialect, or the Southern Teochew dialect, retain much of the features of, and you will really feel like you're studying the mother language.

> Some researchers have been known to use Korean pronunciations of words to help triangulate to what ancient Chinese characters would sound like

That is rather an exaggeration. Korean and Japanese can be used as guides for Middle Chinese pronunciation, but they aren’t all that useful compared to using China’s own rhyme tables and running the comparative method on the Chinese dialects. For stages of Chinese before Middle Chinese, neither Korean nor Japanese are useful at all.

>Japanese students learned Korean even quicker than the Chinese students - in addition to using Chinese words, Japanese and Korean have the same grammar structure ( Subject Object Verb ) whereas Chinese and English are different (Subject Verb Object)

I've never found this a convincing reason as to why a language is difficult to learn. Word order is learned once while vocabulary, speaking and listening need to be learned and refined consistently over many years. All of these become harder the further the target language is from the ones you already know. Would French really be significantly harder to learn if they said 'Je le ballon frappe'?.

I also think this applies to some other 'hard' qualities of languages, like pronouns and alphabets (i.e. not logograms). Conversely, freedom in how you approach grammar in a sentence makes it harder, not easier, because while you're able to structure a sentence how you like it makes it harder to know what sounds idiomatic and it increases the number of forms you need to listen and read for.

Tangential thought: one of the reasons that vocabulary is annoying can be presented in CS terms. Imagine reading a sentence like 'Do you want to foo the bar?'. From the context you can probably understand what kind of words foo and bar are but you don't know what they mean. Yet they're the most important words in the sentence. This isn't a coincidence: the least common words in a sentence tend to be those with the highest entropy. If there was another word that carried the same meaning but was more common, the author would be more likely to use it. So it's not enought to learn 90% of the words, you need the niche words too.

A different word order, any different word order, changes language learning from "learn a bunch of substitutions for my native language" to "learn a completely new system". There's a huge mental overhead for the latter.

The effort of learning a different grammar is dwarfed by the effort of memorizing a sufficiently large dictionary to understand everyday speech. To read at an elementary school level you need 5-20k words.

I thought that too, but as a bilingual person who studied Latin for several years... either I'm linguistically handicapped, or the grammar was legitimately tough to wrap your head around. Or both. Some of it was that I really did struggle with the grammar. Some of it was that sometimes we really were reading syntactically complex writing.

If you're a master of the grammar, I think it's possible to start figuring things out from context once you've learned a critical mass of words

Yes, but "critical mass" is much larger than you'd commonly think. It's the low frequency words that contain most information in a sentence.

> It's the low frequency words that contain most information in a sentence.

This is an incredibly insightful point.

Thank you. That's a very succinct way of putting it. I wonder if there's a term or concept for this in linguistics.

That's not true. People usually learn around a thousand words per year. Someone in their 20s is expected to have a 20k vocabulary but much of that is academic or professional vocabulary. i.e. the overlap between a mechanical and a software engineer might only be 15k words.

Is that number for active or passive vocabulary? For active, it seems reasonable, but GP is talking about reading comprehension, i.e. passive vocabulary.

There's a lot more to learning a different language than learning substitutions.

For starters, there is hardly any bijective mapping between words. Translating 'run' to German could mean any of 'rennen, renne, rennst, rennt, renn, laufen, laeuft' and I am sure I am missing quite a bit.

And that only scratches the surface, let's not get started about Grammar or idioms, etc.

The word order really only adds negligible overhead.

You get closer to something bijective when you only consider lemmas, but then you'll probably have around 1/3 that are bijective, 1/3 that 1->2 or 3 and 1/3 that are 2 or 3->1.

So learn it. You'll get to practice it with every sentence you ever use.

Conversely, you'll only use 'ballon' when you take about un ballon. I hope that's frequently enough for you to remember the word & its gender. At least it looks like the word 'ball', something you don't get with Vietnamese's 'bóng' or Chinese's '球[1]'

And you'll still get funny looks if you say you're getting ready to go to the New Year's ballon, because substitutions fail with homonyms in both directions. And you'll have to learn and relearn about ten thousand of these substitutions.

[1]maybe?

More like building a completely different persona for each family of languages. “Learn a bunch of substitutions” approach only goes so far, your language specific part of identity must be separated at where thought processes for languages diverge and necessary new parts must be built.

Philosophers deny linguistic determinism(a notion that language sets the ceiling for thoughts) but that’s just talking Turing Completeness.

I think you're hinting at the fact that all language acquisition ultimately boils down to vocab acquisition for the long tail to mastery.

As for why people complain about vocab, it's probably because the dropout rate for language learning is huge (look at the completion rates on duolingo or popularity of beginner resources vs intermmediate vs advanced). By the numbers I suspect most learners are struggling with grammar, and most will drop out before "finishing" with grammar and being left with the long tail of vocab.

This has been my experience that in the long run everything else is a distant second to vocab acquisition. I think even within vocab the long tail you have to worry about is overwhelmingly nouns. Verbs are also numerous but can often be inferred from context, replaced by simpler verbs etc. 95% of the time when I’m watching tv news and can’t follow its because they use some finance, politics, medical terms etc that cannot be guessed. There are simply far more ideas/things in the world than anything else. Unfortunately the truth that language learning is a long slow grind to learn nouns doesn’t fit well with those who want to make it a more intellectual exercise or other pedagogical goals of academic settings.

Just learning the words sounds great until you find out that Korean is agglutinative, and Koreans don't even agree on where the word boundaries are.

> Tangential thought: one of the reasons that vocabulary is annoying can be presented in CS terms. Imagine reading a sentence like 'Do you want to foo the bar?'. From the context you can probably understand what kind of words foo and bar are but you don't know what they mean. Yet they're the most important words in the sentence. This isn't a coincidence: the least common words in a sentence tend to be those with the highest entropy. If there was another word that carried the same meaning but was more common, the author would be more likely to use it. So it's not enought to learn 90% of the words, you need the niche words too.

If you don't understand what a word means, can't you ask the other person?

Not in every situation & not if you're not good enough to understand the explanation.

That's an impossible bar to clear, considering that it happens to native speakers of any language too.

Subject object verb is the least of your problems if you're learning Japanese. Moreover, English supports that order: it's just not the canonical one. So it is not that alien. "Yoda speak" appears in sentences like "In god we trust".

Japanese also deviates from its canonical order at times, by the way. For instance, sometimes the topic is extraposed to the end of the sentence, even after a supposedly sentence-ending particle like よ. E.g. "ばかだよ私は". Sometimes that can be because the speaker realizes mid-sentence that it requires a topic to be clear, and you can tell when this is the case when the topic is preceded by a pause. But it has a nuance of its own, so it is used deliberately, even in writing.

"Yoda speak" is actually object-subject-verb, not subject-object-verb. (And the sentence "In god we trust" is also OSV order). Subject-object-verb is not really present in English even in archaisms or other weirder contexts.

It's not absolutely unknown: "With this ring, I thee wed" from the Book of Common Prayer marriage service seems to me to be SOV (with an adverbial bit in advance), though it's definitely archaic. I imagine it features a bit in poetry too.

The word order in that sentence is less important, because it uses old pronoun "thee" as opposed to "thou" (analogous to German "dich" vs "du"), thus indicating the subject and object.

"With this ring, thee I wed"

"I wed thee, with this ring"

...

"Thee, with this ring, I wed" (???)

When you use "you", suddenly there is only one obvious order.

You, with your attitude, I cannot accept!

I think your example highlights one of the amusing things about Japanese. That is the complete omission of the subject. As you said, the 「私は」 might only get thrown on as an afterthought. You can have entire conversations without ever actually voicing the subject. It can come off like a British comedy sketch for somebody to come into the middle of an exchange and try to figure out the topic from the last half of the conversation.

Kind of like using genderless pronouns in English takes away some hint to help you catch on. Only there's no pronoun at all. For English speakers it can get even worse when all that is spoken is a succession of passive verbs. Hilarious.

Edit: I should say that I never noticed a native Japanese person to be confused about the unspoken topic. They seem to be able to intuit it just fine in any context. I'm mostly just laughing at my own inability there.

> Edit: I should say that I never noticed a native Japanese person to be confused about the unspoken topic. They seem to be able to intuit it just fine in any context. I'm mostly just laughing at my own inability there.

Sometimes they will get confused and clarify with a question: 私って？誰のこと？私とは、どちら様のことでしょうか？

Just as a note, your sample Japanese seems odd to me. I'm way out of the loop on anything modern, but it is weird to see written Japanese with non-Japanese punctuation throughout.

Interestingly, it has the effect on me of giving it vocal inflection which works quite well in the context of the example.

This is a great comment, it's something I have never really thought about, but now that you mention it...

、is "correct" Japanese and is used in formal or serious communications.

？ isn't technically correct in Japanese but it's become pervasive in informal to semi-serious communications in the last 15-20 years, but particularly the last 10 years.

It's comparable to the "passive aggressive full stop" in English, and the shift has happened over roughly the same time period too. These days, seeing someone write か。 instead of か？ makes me think they are either old or upset*

*in anything other than quite formal keigo mails (JMO, I am mid 30s)

> Interestingly, it has the effect on me of giving it vocal inflection which works quite well in the context of the example.

It's interesting that you write that. I would say that is why people write that way:

終わったのかもしれませんよ？

終わったのかもしれませんよ。

In [0] there is a chart that shows that Heisei-born are almost as likely to exclusively use ？ to indicate a high pitch at the end of a sentence as 40s+ are likely to exclusively use 。

[0]https://www.nhk.or.jp/bunken/research/kotoba/20170401_4.html

>Learning Chinese is like learning Latin before learning a Romance language.

>The added benefit is that Chinese isn’t a dead language.

I wouldn't go that far. It's more like learning Greek and hearing many of the same root words, sounds, etymology, etc. in English.

An Italian speaker can listen to Latin and grok maybe 25%. A Japanese speaker can listen to Chinese and get maybe 2%, max.

Words like "library" use the same Chinese letters, and have very similar pronunciations, but that's about the extent.

I'd say the best way to learn would be Japanese, Korean, then Chinese (what I did).

Source: I speak Japanese, Korean, and Chinese.

That sounds similar to English and German. There are a few words that are still somewhat connected, like water / wasser, but knowing one gives little insight into the other.

There's also "Gift" (Poison), "Pickel" (Pimple), "Wo" (where), "Wer" (who), "Fahrt" (travel), and so on...

> I'd say the best way to learn would be Japanese, Korean, then Chinese (what I did).

Why? I speak Chinese and used to be fluent in Japanese. I find my Chinese (and English + katakana) a huge help when visiting Japan in recent years.

What's your rationale for this ordering?

Because when you learn Japanese, you have to learn kanji, the more traditional strokes (non-simplified), both Japanese and Chinese sounds, and the grammar is closer to Korean (and closer pronunciation for a lot of words).

So, it's better to learn Japanese, quickly pick up Korean, then learn Chinese.

Chinese last since you have to learn simplified characters, the grammar is different, tonality is quite different, etc.

> kanji, the traditional strokes (non-simplified)

Uh, both Japan and Chinese use different simplified glyphs. And they overlap a lot because they are different standardizations of handwritten glyphs in the bigger Sinosphere. I would be surprised if you indeed learned the traditional characters while learning Japanese.

>Uh, both Japan and Chinese use different simplified glyphs.

Wrong. Though there are divergences, Chinese in mainland China use simplified Chinese characters (https://en.m.wikipedia.org/wiki/Simplified_Chinese_character...).

Kanji is more closely related to the traditional Chinese characters (https://en.wikipedia.org/wiki/Kanji#Local_developments_and_d...) than the mainland simplification.

>I would be surprised if you indeed learned the traditional characters while learning Japanese.

Kanji is closer related, or exactly the same as traditional Chinese characters (used in HK, Taiwan, etc.) than simplified.

My understanding of Chinese (the language) is limited, but having learned Hanja (essentially traditional one) first then learned Japanese later the divergence was already significant. Or, more quantitatively, the Joyo Kanji has 364 simplified characters out of 2,136 (~17%) [1] while the mainland China simplified probably about 2,000 out of 7,000 common characters (~28%) [2]. This divergence is not the biggest deal, but still big enough to refute the claim that Japanese is easier to learn because it's "non-simplified" (there may be other valid reasons though).

[1] https://en.wikipedia.org/wiki/Shinjitai#Simplifications_in_J...

[2] https://en.wikipedia.org/wiki/Simplified_Chinese_characters#...

>Or, more quantitatively, the Joyo Kanji has 364 simplified characters out of 2,136 (~17%) [1] while the mainland China simplified probably about 2,000 out of 7,000 common characters (~28%) [2].

It's 17% and 28%, but much, much higher for the most commonly used characters, which is where the divergence between simplified and traditional grows even more.

>This divergence is not the biggest deal, but still big enough to refute the claim that Japanese is easier to learn because it's "non-simplified" (there may be other valid reasons though).

I didn't make that claim. Maybe your talking about another thread?

It's not that it's easier to learn, it's that the path of going Japanese -> Korean -> Chinese is the easier, most logical path. The diverging simplified Chinese standard characters being a large reason behind that.

Doesn't the order matter only if it's harder to do this the other way around?

Anecdotally for me, learning Chinese first with simplified characters was totally fine, and without formal study, I can read traditional characters fine 90% of the time and guess a remaining ~8% of characters I don't immediately recognize for a total ~98% comprehension, kind of like figuring out how to read medieval gothic font.

Knowing Chinese, learning Japanese and its own simplified characters (shinjitai) was a breeze compared with my classmates who struggled especially with kanji, and also pronunciation. Until I compare hours with a native Japanese speaker learning Chinese, it's hard to tell if learning Japanese, then Chinese would have been easier or not.

> Anecdotally for me, learning Chinese first with simplified characters was totally fine, and without formal study, I can read traditional characters fine 90% of the time and guess a remaining ~8% of characters I don't immediately recognize for a total ~98% comprehension, kind of like figuring out how to read medieval gothic font.

AIUI that's not a recommended approach; my friend is learning Chinese and his teacher was insistent that he should learn the traditional characters first because it's much easier to pick up the simplified characters from the traditional ones than the other way around.

It's definitely true that it's easier for readers of traditional to learn simplified than visa-versa. After all, the simplified script was created by people who were already using traditional and were attempting to make logical and consistent simplifications!

On the other hand, your friend might find some very common traditional characters brutal to learn as a beginner. 醫生 (doctor) often comes up within the first few lessons and 鬱悶 (depressed) within the first couple of semesters. In contrast, someone starting with simplified won't be so daunted learning the simplified equivalents 医生 and 郁闷. Later, as an advanced student, the traditional forms aren't so daunting as they would have been to start with.

It's still more work for someone who knows 医 to learn the rest of 醫 than it is for someone who knows 醫 to learn that they just need to write the first 8 strokes that make up its upper left corner to make the simplified version, of course.

It's an interesting point--I underestimated the difficulty of learning to handwrite characters. Knowing simplified I can read traditional pretty effortlessly and even communicate with a relative in HK typing in traditional, but cannot write it by hand from memory. I liken it to being able to read but not draw the Gothic "T" in the New York Times.

Still, I'd claim that regardless of the direction, learning the simplification map (and some patternless exceptions) is a small incremental task compared with the thousands of hours learning the broader language. What's a few hundred mapping rules + exceptions, after learning 3000+ characters? If someone asked me which character set to start with, barring real issues with availability of education material, I'd say to learn whatever you want if you're committed to it! Just be ready to spend a few hundred more hours at least due to the increased amount of information to memorize in traditional.

You can definitely get a deeper appreciation of a language and get extra decoding tools for learning related languages when you learn about the history and etymology of words.

As a romance language speaker, knowing about greek and latin particles helped me greatly in learning english of all languages, just because of the sheer amount of appropriations english does.

Something that sticks with me from my early japanese learning after all these years is that, for example, "あ" (japanese "a") is a stylization of "安" (pronounced "an"). This helps make an association similar to how some kids books overlay hiragana over pictures of things that start with that letter, e.g. overlaying "い" (i) around a strawberry (ichigo). There are tons of similar examples in both japanese and chinese (pronunciation prefixes being a common tool used by chinese learners)

Speaking sure, because both Korean and Japanese were fully formed languages that borrowed a writing system. I've come to view Kanji/Hanja as an information layering on top of the language used to increase the information density/clarity of the written languages.

If you look at written Japanese/Korean (with Hanja) and you get a much higher transfer. Esp. Someone who knows Chinese trying to get by in Japan.

> Learning Chinese is like learning Latin before learning a Romance language. The added benefit is that Chinese isn’t a dead language.

As a native Korean speaker this is an apt analogy, even more so because it won't help much initially but will help a lot once you've got to the later stage of learning. This is of course a valid strategy if you want to eventually learn both Chinese and Korean.

Yes - you should learned the language that has meaning to your life. If you don’t have a reason to use the language you will forget it all.

And agreed - once you get into more “college level or technical” Korean for lack of a better word, knowing the hanja/ Chinese characters makes vocab acquisition easier.

> The Chinese pronunciations have changed. Some researchers have been known to use Korean pronunciations of words to help triangulate to what ancient Chinese characters would sound like (in combination with poems where you know what rhymes to expect etc)!

Modern Korean isn’t a tonal language tho, the length of the vowel changes the meaning. So how does that work?

Modern Korean alone definitely isn't sufficient to reconstruct from, but including it in an analysis can add some information that isn't preserved in other branches.

Chinese itself wasn’t much of a tonal language until a little more than one thousand years ago. Modern Chinese tones emerged in the Middle Chinese era as a way to avoid homophony/preserve distinctions of meaning as a lot of old final consonants were lost.

Korean does not have vowel length change the meaning, however Japanese does.

Korean used to be tonal, and you are right it now longer is.

Korean is used in combination with other data points to determine pronunciations.

This works for old English plays, for example too. You can use poems that should rhyme, plus commentary from linguists at the time, to guess what certain words would sound like “back then”.

Just saying that Korean is another super rare but fun random fact that I have heard of being used at least once in combination with ancient Chinese poems, to guess at the old Chinese pronunciation.

https://en.m.wikipedia.org/wiki/Korean_phonology

> In 2012, vowel length is reported almost completely neutralized in Korean, except for a very few older speakers of Seoul dialect,[14] for whom the distinctive vowel-length distinction is maintained only in the first syllable of a word.[13]

So it does have vowel length change meaning but it’s almost kinda died out like tones?

Native Korean speaker here, born in the 70s and grew up in Seoul. We all learned a dozen short-long pairs at school (like 눈 nun being either eye or snow), and of course the education got all of us tricked into thinking that we're making these distinctions (if only we try hard enough). But looking at it objectively, I don't think a single person around me used vowel length distinction. If you have to "think" whether a word is long or short, that means the distinction is dead. No English speaker has to "think" whether fan or pan is the correct word. An actual sound distinction comes automatically for native speakers.

Similarly, ㅐ/ㅔ ("ae"/"e") distinction was dead before I was born, but schools insisted that they're well and alive. Not sure what they're teaching these days.

I can't speak Korean, only a little bit of Mandarin :D so I'm really happy to have your reply, thank you! Makes sense!

Vowel lengthening matters in Korean technically but is more often context-sensitive than length sensitive like it regularly is in Japanese for reasons I don’t know. 눈 can mean snow or eye in Korean depending upon how long you extend the vowel as recently as 1970, for example, but many Koreans (especially younger Koreans) have less sensitivity and don’t use it. When I learned Japanese there were more words where it was easier to confuse them and thus require a distinction, probably preserving length.

Overall Korean historically has had a tendency toward reducing tones and sounds (it used to have f and z sounds like in Chinese, for example). It continues to eliminate vowels where a lot of Koreans are not pronouncing the sound 최 fully and eliminating lip rounding to make it simply 체.

> When I learned Japanese there were more words where it was easier to confuse them and thus require a distinction, probably preserving length.

I did this to unintentional comedic effect the other day by referring to 「星の王子さま」as「星のおじいさま」.

My spouse is Chinese and I tried to learn Chinese for obvious reasons. However, they speak multiple dialects in practice such that it's hard to focus, and there are limited training materials for some of the dialects. Generally there is Cantonese, Mandarin, and their local dialect, which is what they use most often.

Same problem here (different dialect though). Let me know if you found a solution.

The English and Chinese Wikipedia articles about the dialect will usually have at least some useful information e.g. about the phonology and maybe links to other resources.

You can also try Wiktionary: https://en.wiktionary.org/wiki/Wiktionary:About_Chinese#Abou...

https://xiaoxue.iis.sinica.edu.tw/ccr/ has IPA pronunciations for single characters across many different dialects.

https://forvo.com/languages-codes/ may have some recorded vocabulary.

https://github.com/laubonghaudoi/Chinese_Rime has input methods for many dialects that can also be abused as a dictionary (associating romanization with Chinese characters you can look up).

https://xefjord.wixsite.com/xefscompletelangs/courses#comp-k... may have flashcards if you're lucky.

You'll probably also want to make your own flashcards. https://github.com/ppwwyyxx/wechat-dump is helpful for getting voice messages out of WeChat.

If you want to practice by passive listening, you can try http://phonemica.net/ , searching on https://youku.com/ or check whether there's a local TV station that has programs in dialect.

Reading material will be almost impossible to find, but maybe there's a bible translation from the 19th century or something.

Also check whether there's a local language preservation group.

Finally, https://zhongguoyuyan.cn/ is supposed make a lot of material available to the public soonish (right now I get a certificate error in Firefox, but you can see the landing page in a less strict browser).

Your spouse must be fairly old school to be using a dialect that isn't mando/canto...

Most millennials I know from mainland china and even Malaysia/Singapore will speak putonghua with their parents/family now. Ofc people from Sichuan and Shanghai love to preserve their dialect. Plus you got the cantos and hokkien/teochew/hakka and even changsharen.

In any case, my recommendation would be to stay with Mandarin and go deep on it. Learning a dialect afterwards is quite straightforward and you can be 80-90% proficient esp with understanding by just developing a sound mapping of each character from dialect->mandarin.

The common idioms you'll pick up with enough exposure.

You'll never regret learning Mandarin first - you'll be able to read/listen to the news and your relatives will understand you. Whereas if you try to learn dialect first you'll find it harder due to the less set of materials AND less people to talk to.

A lot of chinese people joke that they learned their 2nd/3rd... dialect by doing KTV and there's definitely some truth in that

I came across this quote from the Mandarin wiki:

> The Chinese have different languages in different provinces, to such an extent that they cannot understand each other.... [They] also have another language which is like a universal and common language; this is the official language of the mandarins and of the court; it is among them like Latin among ourselves.... Two of our fathers [Michele Ruggieri and Matteo Ricci] have been learning this mandarin language... — Alessandro Valignano, Historia del principio y progresso de la Compañía de Jesús en las Indias Orientales, I:28 (1542–1564)

I always find the claim that dialects in China vary so much that people cannot understand each other very western-centric. As a native speaker, there are shockingly few dialects that are completely incomprehensible. Someone from Shanghai will be able to speak to someone in Chengdu with their own dialect, just slower.

The claim may be misguided or wrong perhaps, but how is it Western-centric?

Take Cantonese (or heck, Teochew or Hakka). You really think this language is reasonably mutually intelligible with Mandarin? Or Shanghainese?

I'm a (non-native, formerly fluent but now pretty poor) Cantonese speaker and I'd describe the difference between Cantonese and Mandarin as roughly the difference between Portuguese and Romanian. Sure, they're both romance languages, and maybe with a lot of work and hand-waving and drawing characters on your hands with your fingers you can get your point across to someone who speaks the other language, but in no way would I call them mutually intelligible, even by talking "slower".

Cantonese and Mandarin have very different pronunciation, vocabulary, and even grammar for common cases. In fact there are extremely common words in Cantonese for which there is no modern chinese character: instead roman words or even single letters (like "D") are substituted in comic strips etc. So you can't even write them down in Chinese for a Mandarin speaker to puzzle through with a dictionary.

I'd say that your perception of these things as "western centric" -- they are not at all -- strikes me as very "northern Chinese centric" view of China. :-)

It's not necessarily a northern Chinese centric view of China as it is a Mandarin (as in the language family) view of China. The vast majority of Chinese dialects in a large band from Heilongjiang to Sichuan (and a bit out to Qinghai and Xinjiang) are more or less mutually intelligible. That mutual intelligibility drops off very quickly outside of that band. In particular a lot of Southeast dialects (Yue, Wu, Min, etc.) fall outside of that zone. But it covers the majority of Chinese speakers. (However, I disagree with the original claim that a Sichuanese speaker could understand Shanghainese).

See e.g. this video: https://www.youtube.com/watch?v=Ps7_NnkL-oM where for almost the entire video a couple is explaining words in their respective dialects (Sichuanese and Hunanese) to one another in their own dialect (i.e. not using Standard Mandarin at all in their explanations). The first half of the entire exchange is fluent and understandable by both people and also a Standard Mandarin audience. Only in the second half when they intentionally start quizzing each other on words they know will be tough for the other person and without additional context do they run into trouble (and occasionally resort to Mandarin).

Also, comparing it to Romance languages masks some complexities of the relationship. https://news.ycombinator.com/item?id=16844074

Heh, this is the perfect place for me to jump in with some movie trivia.

Have you ever watched either Ip Man or "The Bridge" (Danish/Swedish cop show)?

In Ip Man, there's a gang of Mandarin speakers who show up in Ip Man's Cantonese world. They somehow are able to speak to each other, and I find it unlikely. Perhaps it's because my Cantonese is mediocre. But it just doesn't sound like you could have a casual conversation across the languages.

A similar thing happens in The Bridge. There's two cops solving a crime on the bridge connecting Denmark with Sweden, and they just talk to each other in their own languages. My guess is most people who didn't study the other language would not know all the slang, as importantly the simple expressions. Though I guess it could be learned easily; I can read a Swedish newspaper, just not process it fast enough to have a fast conversation with a colleague.

Similarly with German and Swiss German. Something about the way the sounds are different makes processing the other one a bit slow if you're not used to it.

> In Ip Man, there's a gang of Mandarin speakers who show up in Ip Man's Cantonese world. They somehow are able to speak to each other, and I find it unlikely. Perhaps it's because my Cantonese is mediocre. But it just doesn't sound like you could have a casual conversation across the languages.

So the mixed Mandarin / Cantonese conversations actually do happen. My wife's family are all native Cantonese speakers, but her uncle married into the family from another part of China. He understands Cantonese perfectly but is more comfortable speaking Mandarin; and everyone else understands Mandarin to various degrees but prefer to speak in Cantonese. So he just speaks Mandarin while everyone else sticks to Cantonese; and it's quite a fluid back-and-forth.

That said, for that to happen so naturally the way it does in the Ip Man movies is really unrealistic. First of all, Mandarin wasn't as well-known in Guangdong in the 1930's as it is now; and there's no way someone from another part of China could just show up and magically understand Cantonese: It would take at least a few months of immersion to be able to "listen" fluidly enough to have that sort of a mixed conversation.

Even more unrealistic is Ip Man 4, where he shows up in San Fransisco, and starts talking in Cantonese to people who have immigrated from Beijing -- including one of the character's 12-year-old daughter who grew up in San Fransisco. Yeah, that's not going to happen.

[Minor grammar edits.]

That's kinda interesting that he can do that. With a bit of recent Mandarin schooling I can see the langauges are quite similar, but as a kid whenever we had a Mandarin speaker visiting I was baffled. Same thing happened when I learned German, as soon as I learned the initial things like the pronouns everything clicked.

It's probably the case that all the often used words that you use in everyday life have changed, but once you get over the barrier there's a pretty smooth learning zone.

Just to be clear, her uncle has now lived in HK for several decades; I'm sure there was a several-month learning curve being able to understand Cantonese. It's not like Hindi and Urdu, where (I'm told) people with no previous exposure to the other language can communicate pretty readily.

But there are a huge number of words which are simply pronounced a bit differently; and the differences in pronunciation have a lot of predictable patterns. So once you're familiar with these patterns, suddenly a Mandarin speaker can understand a huge amount of Cantonese when spoken in context.

Here are ten Shanghainese sentences, four short ones and six long ones:

1. https://audio.tatoeba.org/sentences/wuu/488406.mp3

2. https://audio.tatoeba.org/sentences/wuu/485590.mp3

3. https://audio.tatoeba.org/sentences/wuu/485698.mp3

4. https://audio.tatoeba.org/sentences/wuu/488621.mp3

5. https://audio.tatoeba.org/sentences/wuu/489046.mp3

6. https://audio.tatoeba.org/sentences/wuu/489736.mp3

7. https://audio.tatoeba.org/sentences/wuu/492871.mp3

8. https://audio.tatoeba.org/sentences/wuu/496402.mp3

9. https://audio.tatoeba.org/sentences/wuu/485676.mp3

10. https://audio.tatoeba.org/sentences/wuu/488252.mp3

They're not especially slow, but you can replay them as often as you want or even slow them down, so I think it should be easier than for someone in Chengdu trying to understand a Shanghainese speaker in person.

For comparison, here are the Mandarin equivalents of those sentences, pronounced by the same speaker:

1. https://audio.tatoeba.org/sentences/cmn/332870.mp3

2. https://audio.tatoeba.org/sentences/cmn/332436.mp3

3. https://audio.tatoeba.org/sentences/cmn/332568.mp3

4. https://audio.tatoeba.org/sentences/cmn/333012.mp3

5. https://audio.tatoeba.org/sentences/cmn/333070.mp3

6. https://audio.tatoeba.org/sentences/cmn/333151.mp3

7. https://audio.tatoeba.org/sentences/cmn/333392.mp3

8. https://audio.tatoeba.org/sentences/cmn/333584.mp3

9. https://audio.tatoeba.org/sentences/cmn/332549.mp3

10. https://audio.tatoeba.org/sentences/cmn/332788.mp3

And here they're written down, together with translations into various other languages:

1. https://tatoeba.org/cmn/sentences/show/488406

2. https://tatoeba.org/cmn/sentences/show/485590

3. https://tatoeba.org/cmn/sentences/show/485698

4. https://tatoeba.org/cmn/sentences/show/488621

5. https://tatoeba.org/cmn/sentences/show/489046

6. https://tatoeba.org/cmn/sentences/show/489736

7. https://tatoeba.org/cmn/sentences/show/492871

8. https://tatoeba.org/cmn/sentences/show/496402

9. https://tatoeba.org/cmn/sentences/show/485676

10. https://tatoeba.org/cmn/sentences/show/488252

If you want some more examples, this search returns a few hundred: https://tatoeba.org/cmn/sentences/search?query=&from=wuu&to=...

To be fair these translations slightly exaggerate the difference between Shanghainese and Mandarin. The speaker is speaking in a more formal register in her Mandarin samples than in her Shanghainese samples which causes greater vocabulary and structural differences than if they were at the same register. In fact I'm a little surprised she doesn't use exact analogs when translating between the two, e.g. 搿就是讲 -> 这就是说 instead of her choice of 这意味着, 越来越结棍了 -> 越来越历害了 instead of her choice of 愈演愈烈, and 开始寻被偷脱个物事了 -> 开始找被偷掉的东西了 instead of her choice of 开始找被偷物品了, since it is both a more accurate translation and reflects the structural similarity of the two sentences much better (there is a one-to-one correspondence between each word of both sentences).

Regardless I certainly would not view Shanghainese as mutually intelligible with Mandarin.

> In fact I'm a little surprised she doesn't use exact analogs when translating between the two

If you check the logs on the sentence page, you'll see that the Shanghainese sentences are actually translated from the Mandarin. So she was going from relatively formal Mandarin to more informal Shanghainese, probably because that's what she's accustomed to speaking.

I didn't really intend these as a demonstration of how different Shanghainese and Mandarin are even when you translate as literally as possible, but more as a way for people on HN who know Mandarin but haven't had much exposure to Shanghainese to try and see how much they understand. I like to think I understood a few words like 自己 and 警察, but maybe that's because I looked at the text first.

This is a super high quality post - thank you!

I think you've also answered my questions in this space before, what kind of work do you do? You seem very familiar with asian languages deeper than just a native speaker/hobbyist might be.

I thought recording 6 was vaguely mandarin sounding, if you squinted you could more or less guess what was being said. Agreed that the rest are hard to understand but I don't think it takes native mandarin speakers much effort to get to a point where they can get proficient in another dialect. IMO learning Shanghainese from Mandarin will be easier than learning Portuguese if you speak Spanish

> Someone from Shanghai will be able to speak to someone in Chengdu with their own dialect, just slower.

This is definitely not true (if they were to use their own dialect).

I'm from the same region surrounding Shanghai and although there are some intelligibility between dialects, it is hard if not completely impossible to have a conversation even between people from adjacent towns. See https://en.wikipedia.org/wiki/Wu_Chinese

I'm a bit surprised about your claim. Anyone from China with a sense of the dialects of the south would know this.

Mandarin speakers ability to understand slowed down Shanghainese is likely a recent development. This is because much of traditional Shanghainese diction, grammar, and even pronunciation has been reduced or morphed to better suit the newer generations of speakers whose principal language is Mandarin.

written cantonese is almost the same as written mandarin.

the problem is that spoken cantonese is not much like written cantonese. it's a combination of pronunciation, accent, idioms, and sentence structures. best advice is for speakers of mandarin or cantonese to learn the other language like a new one. (there's also the written form of spoken cantonese, written spoken cantonese, if you will, that is the written representation of spoken cantonese using a set of special characters of sinitic origin).

if you are from guangdong you will learn all systems of language - written mandarin, spoken mandarin, written cantonese, and spoken cantonese. some hongkongers don't ever learn spoken mandarin correctly, getting them to speak mandarin in front of a camera is always a source of good laughs.

those who argue that cantonese are mutually intelligible with mandarin will typically use cantonese newspapers and official documents as evidence. those who argue that they aren't will point to spoken and colloquial cantonese as evidence. the objective point of view is to accept that cantonese as we know it exists as two separate languages used for two different purposes.

it's true but also largely moot. mandarin is enough, nearly everyone speaks it.

If you're learning Chinese from her, you might be interested in my project for couples learning each other's languages! https://learncoupling.com

My university Japanese class was full of Korean and Chinese international students. It was hard to keep up.

Reply on news.ycombinator.com