Hacker News new | ask | show | jobs
by amyjess 1391 days ago
Japanese grammar is starkly minimalist. It's not hard to learn at all, the basic structure is almost purely agglutinative, and the word order is consistently head-final in all cases (e.g. SOV for sentence and modifier-modified for not only adjectives but also relative and appositive clauses), and it helps that Japanese doesn't grammatically track several things that other languages do, such as person, number, or gender.

There are only two real problems:

1. The writing system is ridiculously complex, and even if you just vow to only write in romaji you also have to deal with the problem that kanji acts as a huge source of both puns and compound words. You can invent new compound words just by jamming together the on readings of a couple of kanji and most Japanese people will understand you. It's also not unheard of in, for example, songs, to pronounce a word one way when singing but write it in the official lyrics sheet using kanji that's normally associated with a completely different word. The closest I can compare to this in other languages would be like if you were talking and using sign language at the same time and you were deliberately signing different words than what you were speaking in order to add subtext.

2. Because a) so many features aren't grammatically tracked and b) Japanese is aggressively pro-drop, a lot of sentences are extremely ambiguous without context. For example, you often can't tell just from hearing the words if someone is saying "I go", "you go", "they go", "he goes", or "she goes" (in Japanese these are all just iku/ikimasu... unless you're going out of your way to put a pronoun in there, but most people don't); you have to parse the sentence in the context of what else is being said in the conversation or by what's going on around you.

4 comments

Kanji is indeed complex but you can memorize Katakana and Hiragana in a couple of hours and they will help you tremendously in things like navigating the public transport signage. They have a few more characters than 26, but each letter has exactly one pronunciation. None of that context-dependent pronunciation stuff like English is full of.

Tackle Kanji after mastering Katakana and Hiragana.

You can memorize hangul (Korean alphabet) in a couple of hours.

It took me a few weeks to be able to read a hiragana sentence, just read not necessarily understand, without needing to look anything up.

> The writing system is ridiculously complex

Most of the complexity is merely that it requires a lot of memorization, though. One has to literally remember a couple thousand of characters, memorizing their meanings and a few possible readings. Obviously one remember them not as a random opaque pictogram but by mentally splitting it in smaller graphemes. And there are, IIRC, like, about 30 or so most common ones that are enough for most characters one would normally encounter.

Either way, it's surely much more complex than systems that only have alphabets or syllabaries, but in a personal (and biased, because I know one and failed to grok the other) opinion some segmental scripts like Arabic are much harder to process.

You're forgetting the difficulty of learning the elaborate system of honorifics, without which you'll be unable to talk to a native speaker without insulting them. The title of this book gives some idea: https://www.amazon.com/exec/obidos/ASIN/4770016247/ref=nosim...
While that is technically true standard 敬語/honorifics (です/ます and a few word choices) aren't really that complicated.

There are additional levels of honorifics which can be far more complicated but (outside of workplace honorifics -- which you will need to practice if you will work at a Japanese company) native speakers usually get some kind of training in how to speak in that exceptionally formal way (the kind of keigo used in restaurants is sometimes criticised for being "incorrect" Japanese and is called バイト敬語 -- usually service workers literally get handed a manual which explains how to interact with customers using this form of 敬語). If you or I had an audience with the queen we would probably also get some kind of training in how to politely speak to her.

Finally, if it's obvious you're studying Japanese and you drop a です or ます the person is quite unlikely to be insulted. Especially if it's not someone who is your superior at work.

Just to relate back to the original point of adopting a universal language, I would guess that if any language were adopted as a world-wide language, then things like honorifics and formal-informal distinctions and gendered articles/nouns would be dropped pretty quickly.
Don’t worry, all Japanese people know you’re a gaijin and won’t be offended.
Oh, they’ll probably be ok if you don’t know the language. But they might be offended that you’re a gaijin.
Lol what
I think the simple grammar makes it harder, it is like assembler has simple syntax comparing to java but it doesn't make it easier to read or write.
You're mistaking simplicity and lack of capabilities.

If a grammar is too limited that it leaves things unsaid and thus ambiguous (well, I can't say that's not a thing in Japanese - it is, sometimes) then yes, it can make things harder because one gotta very carefully watch out for the context to be able to comprehend what's going on.

But if grammar is just simple - e.g. if there are no or almost no irregular verbs (Japanese has only two), or no concept of grammatical gender (so you simply don't have to care about anything related to your nouns and can indicate gender using an extra word if that starts to matter), it's probably just fine. I mean you're not losing anything in that case.

Considering your example (even though I don't think it's a good idea to compare machine and human languages in general) - many assembly languages do NOT exactly map to the actual machine code. Say, many assembly languages remove the concept of argument size which can be compared to a grammatical gender (your operands are your nouns and instruction is a verb). Let's say we're talking classic x86. In Intel syntax you can write "MOV AX, 0" and "MOV EAX, 0" but bytecode would be different - 66B80000 vs B800000000, the assembly had lost the 0x66 operand-size override prefix. There is AT&T syntax that has distinct "MOVW" and "MOVL", but the point is that it's a potentially unnecessary complication that proves to be not needed as everything is pretty much obvious without it.

I understand that simple is better than easy in many cases, that's the whole idea of lisp like languages, but I am not sure it is true for languages, especially for comprehension. Because things sound the same and there is no distinction it makes it hard for listening or even reading comprehension. Many time it is much easier to quickly decipher what a javascript function is doing than lisp because the lisp simplicity doesn't give you those immediate anchors to look at, it all looks the same.