[Bregalad]

I know japanese words can be written with kana and with kanjis, but I wonder what decides whenever you go for one or another. As far as I understand, kana is used if this is a video game on an early/limited system, if this is intended for kids, or if the kanji would be too "compilcated", of if there is no kanji at all in the first place.

So words can be written using both, but is it allowed to mix both within the same word ? For example are 日ほん or に本 legal japansese for 日本 ? (I'd guess no)

What about if you glue multiple words for making a single word ? For instance, to say 日本人, is it legal to write it にほん人 or 日本じん ? (I'd guess yes)
Thanks.

[pmt7ar]

Kanji is used most of the time, whenever its possible to use it. The only reason to don't use kanji is if there is some limitation in doing so:

1) Aesthetics purposes: This is the primary reason for japanese natives, and it comes naturally (you don't think of it explicitly). There is an implicit "ratio" on how much kanji would it be "too much" -i.e. it would look ugly, or make your eyes or head hurt if reading for a long time-. So most of connecting words, particles and adverbs ended up being written more often in kana than kanji. For example: たくさんvs沢山 , もっともvs最も , いつ・どこ・なぜ・するvs何時・何処・何故・為る, -ら・などvs-等・等. Even particles have kanji, の・まで・だけ・とも・ながら vs 乃・迄・丈・共・乍ら. And they are all used, it depends on the author and its aesthetics criteria. You can find もちろんとも or 勿論とも, but maybe 勿論共 its "too much".  Every one has its own "ratio of kanji-kana". Nouns and compounds are mostly kept on kanji, while every else, if possible, in kana.

2) Technical limitations. This is what happens on early systems who couldn't display kanji by either space or resolution limitations.

3) Target: If the target audience is younger, you tend to use kana versions of only very known words. Complex words, however, mostly need to be kept in kanji to understand its meaning. Better than writing in kana, furigana is used to aid reading (a staple on shounen-targeted media).

3) "日ほん or に本". No, please, no. That's horrible and shouldn't be done. It happens however on some games that doesn't have all the kanji data, but it's painfully to read and to understand. に本 is 2本 rather than 日本. Thumb rule, you don't split a single word in kanji and kana. What you can do, however, is split suffixes and prefixes, for example 彼ら (vs 彼等) , 日本じん (vs日本人) and だい統領 (vs大統領), 友人たち vs 友人達.
子ども can be found very often, this is splitting a single word but it comes by reason (3) (aimed towards children who doesn't know 供 yet) or (2) if it's inside games, or because 供 is ateji and somehow it's "permitted". In papers, news, literature and the like 子供 is more common, but in educational material for example, it is 子ども. *ateji is another case when kana used more often: クラブ/倶楽部, タバコ/煙草,  てんぷら/天麩羅, ページ/頁, ビール/麦酒, ロマン/浪漫.

[BRPXQZME]

I’ve seen constructs like 日よう日 in the wild for children’s stuff, but for the most part, in modern Japanese (very old writing gets a pass here) vocabulary past very basic reading is treated as if it’s the same literacy level of the kanji to write them (aside from in some extra-pretentious stuff); either that or the word isn’t usually written in kanji anyway.

I don’t know about having a ratio to kana to kanji; what I can say is that on a word per word basis, there’s usually a general preference to write it with kana or one to write it with kanji (especially in normal, not-super-formal writing; it shouldn’t come as a surprise the preferences change depending on the exact character of what’s being written—common animal and plant names, for instance, are typically katakana in science and kanji in literature). It is highly correlated to part of speech as observed above, though. The dictionaries I use convey this preference through the example sentences. In most cases, you would do well to follow this preference so that people can read what you’re saying. But it’s only a preference and what you’ll usually see. Exceptions abound.

Basically, don’t mixi kana and kanji arbitrarily, even because of technical limitations. It’ll annoy readers.

[Bregalad]

Thanks guys. I have another related question.

Is there a way you can you tell, without knowing much japanese, if character have a weird accent/dialect ?

[pmt7ar]

Thanks guys. I have another related question.

Is there a way you can you tell, without knowing much japanese, if character have a weird accent/dialect ?

Mmm.... nope. You have to know one kind of japanese, so to be able to recognize when something speaks "weird". If it's with audio, maybe you can because you can tell the different rhythm while they speak-

There are words used more in one region like おおきに, ほんま, アホ, but most of dialect differences come from grammar and conjugations. Like や instead of だ, へん instead of ない, different 音便 like ゆうて instead of いって (for 言う).

So at least you have to know some words of a dialect to be able to tell. Kansaiben it's the easiest, it's everywhere.

[Bregalad]

Ah... this suck.

Does characters in Capcom games talk Kansaiben (since Capcom is the only major video game located in Kansai) ?

[BRPXQZME]

Generally only a small minority. When Capcom sells domestically, they sell to all of Japan. The same can be said of nearly all pop media produced in Kansai; you will see this from SNK, Key, Nintendo, KyoAni, etc. There are definitely exceptions (Lovely Complex and Abenobashi, for instance), but these are rare enough that they prove the rule. Outside of certain genres, dialect speakers in most productions are usually either some token character in the cast or fairly unimportant.

Even when characters speak in dialect, writers tend to avoid going out of their way to make something completely unintelligible to the general population unless it’s a joke about unintelligibility.

[Bregalad]

I guess I see, they stick to standard japanese as much as they can, but sometimes leak a few dialect things in unintentionally.

But the same question can be asked reverse-wise, how intelligible is standard Japanese for populations of regions far away from Tokyo ?

[pmt7ar]

Standard japanese is understood everywhere, even in Okinawa and Aomori inaka regions.
Think of a TV news announcer, you can't have a cleaner japanese than that, and that's listened nation-wide.

[Bregalad]

What does it mean when they talk in katakana for words which are not "foreign", not animals, not surnames and not animals (I think) ?

[BRPXQZME]

Usually robots, but really, use of katakana doesn’t necessarily signify anything, just like it doesn’t mean anything if I CAPITALIZE WITHOUT GOOD reason. They are just another set of kana with some conventions, but don’t expect 100% compliance with these conventions. Typical uses of katakana:
- foreign proper names
- loanwords
- plant and animal names in a less literary, more scientific context
- given names (rarely in real life, but a certain amount of characters in fiction)
- slangy terms (e.g. ネタ)
- emphasized words
- to set kanjiless yamatokotoba apart from more grammatical concerns (especially if the ateji suck to write)
- some brand / business names
- onomatopoeia (louder than hiragana)
- loud, brash words in someone’s manner of speaking (children / “big guy” stereotype)
- in kanji dictionaries, on-yomi are in katakana and kun-yomi are in hiragana

Again, these are only guidelines, not rules. And conventions can change; I’ve seen some books from over a century ago where katakana is used as the normal print (it should be noted that in text that old, that is far from the only difference). It’s also done this way in some text for the anachronistically styled Sayonara Zetsubou Sensei, in what seems to be a nod to that convention.

[Bregalad]

Thanks !!

However this only rise more questions.

1) Is there any official "spellig" for load words ? Especially conserning the elongation marks and the sukons, it seems those are placed almost arbitrairly and are very hard to get right (at least I come from a language where the notion of "long vowels" doesn't exist at all and where there is no "germinate consonants" either so this doesn't help).

2) When they see a loan word in katakana how can they know if it's from English or Dutch, Portuguesse, French, German, etc, etc... ? By keeping the original spelling it would be much easier to know which language the word comes from, doesn't it (even if it makes it much harder to pronunce right) ?

3) How common is it to have loan words in hiragana, as seen in the game "Hebereke" (because of technical limitations) ? Can they still "recognise" it ?
 

[BRPXQZME]

1) Is there any official "spellig" for load words ? Especially conserning the elongation marks and the sukons, it seems those are placed almost arbitrairly and are very hard to get right (at least I come from a language where the notion of "long vowels" doesn't exist at all and where there is no "germinate consonants" either so this doesn't help).

The transliterations I’ve seen from a number of languages use fairly regular rules to imitate the original pronunciation; anything different I’ve ever seen seems to either be a really old loanword or be based on some mistake. If you know the rules and the theoretically correct pronunciation under the rules, you’ll almost always arrive at a canonically correct katakana transliteration.

There are, however, some variations for words where the original sounds don’t exist in standard Japanese (バイオリン and ヴァイオリン are both used; the former is still more common). In practice, many Japanese speakers do not produce the sound indicated by these strange, relatively new constructions when attempting to read them aloud.

2) When they see a loan word in katakana how can they know if it's from English or Dutch, Portuguesse, French, German, etc, etc... ? By keeping the original spelling it would be much easier to know which language the word comes from, doesn't it (even if it makes it much harder to pronunce right) ?

Speakers don’t always know and don’t usually care; aside from linguists, most people think in terms of the sounds they use to speak their own language. This is no different from how loanwords usually work in English, for example: most English speakers are completely unaware of the provenance of most of their words, and native speakers of English who attempt to match foreign pronunciations of words that have long been integrated into the language with some monstrous anglicized pronunciation run the risk of sounding like they’re putting on airs. Even if that pronunciation is like nails on a chalkboard.

3) How common is it to have loan words in hiragana, as seen in the game "Hebereke" (because of technical limitations) ? Can they still "recognise" it ?

Not as common as katakana, but it exists. All-same-kana text can almost always be puzzled out after looking at it for a while, although there are usually other hints. Long vowel markers (expected from loanwords by convention) and sounds which exist but are rare in Japanese tend to point towards being loanwords. As a stylistic choice, it usually comes off as a bit childish (which suits that game just fine). There do not need to be technical limitations involved.

In addition to that, there are some words that have been in Japanese for long enough and/or with such common usage that they often get hiragana treatment. Most famously たばこ (in that case, it doesn’t hurt that it sounds like it could be a Japanese word).

[Bregalad]

Mmh, so hiragana when it should be katakana make it "childish", and katakana when it should be hiragana make it "robot". Fascinating !

The transliterations I’ve seen from a number of languages use fairly regular rules to imitate the original pronunciation; anything different I’ve ever seen seems to either be a really old loanword or be based on some mistake. If you know the rules and the theoretically correct pronunciation under the rules, you’ll almost always arrive at a canonically correct katakana transliteration.

Sounds interesting, however considering how few possibilities plain kana offers (without any modern extentions) it's impossible to make anything close to the original pronuncation in many cases. Even with the modern extensions it's still a problem. For example you can't even transliterate "see" without having it turned into "shee".

English has the problem that many volleys are distorted ("a" randomly becomes "e", "e" randomly becomes "i", "i" ranomly becomes "ai" and "u" randomly becomes short "a", most notably) and this is extremely annoying for non native speakers.

In French we have about 3x the amount of vowels that Japanese have so it will alias horribly. (the lack of "v", "f", and a true "r" doesn't help either). Elongation marks and germinate consonants just can't make up the lack of more vowyels ^^

I wonder how they do to transliterate international names/words. For example let's consider "Robert" can be French, German, English (+ others ?) and "Roberto" can be Italian, Spanish Portuguese (+others) ?
So it could be transliterated in several correct ways :
ロベル (french like)
ロベルト (german like)
ロッベールト (italian like)
ウオベート (english like)

[pmt7ar]

Mmh, so hiragana when it should be katakana make it "childish", and katakana when it should be hiragana make it "robot".

Katakana offers a different visual impact, so its often used in something that doesn't sound natural, hence the robot.
It's used also, to difference "live" speaking from a TV, a telephone recorder, the partner on the other side of the line... speaking normal japanese, but with katakana instead of hiragana. People speaking through a megaphone, through a wall, with something in they mouths... it offers a visual difficulty that is used aesthetically to say "this is a pain in the ass to read, same way as the character doesn't hear it perfectly clean".

Also, emphasis. Shouting. IT'S THE JAPANESE VERSION OF CAPS. It isn't 殺してあげる, its 殺してヤル！.

however considering how few possibilities plain kana offers (without any modern extentions) it's impossible to make anything close to the original pronuncation in many cases. Even with the modern extensions it's still a problem. For example you can't even transliterate "see" without having it turned into "shee".

Japanese is more capable than english in this aspect. English pronunciation varies with the words. The "a" in "cat" is not the same "a" in "brain" .
Yes, japanese has fewer sounds as a language, so it's not a problem pronouncing "see" as "shee".
You couldn't write a chinese word in english no matter how hard you try. You can't write a tonal language with a roman alphabet unless you make up a system and the reader is educated enough to understand it.

I wonder how they do to transliterate international names/words. For example let's consider "Robert" can be French, German, English (+ others ?) and "Roberto" can be Italian, Spanish Portuguese (+others) ?
So it could be transliterated in several correct ways :
ロベル (french like)
ロベルト (german like)
ロッベールト (italian like)
ウオベート (english like)

Exactly, the same as any other language. It will be different depending on the language the word comes from.
You write Paul if its in english, Pablo if its in spanish, and Paulo if its portuguese.

[BRPXQZME]

Sounds interesting, however considering how few possibilities plain kana offers (without any modern extentions) it's impossible to make anything close to the original pronuncation in many cases. Even with the modern extensions it's still a problem. For example you can't even transliterate "see" without having it turned into "shee".

There is スィー (but almost nobody uses that in practice).

English has the problem that many volleys are distorted ("a" randomly becomes "e", "e" randomly becomes "i", "i" ranomly becomes "ai" and "u" randomly becomes short "a", most notably) and this is extremely annoying for non native speakers.

It’s not random! It’s... roughly 60% regular

I wonder how they do to transliterate international names/words. For example let's consider "Robert" can be French, German, English (+ others ?) and "Roberto" can be Italian, Spanish Portuguese (+others) ?
So it could be transliterated in several correct ways :
ロベル (french like)
ロベルト (german like)
ロッベールト (italian like)
ウオベート (english like)

Usually ロベール, ローベルト, ロベルト, and ロバート, respectively from what I’ve seen (ロベルト is also used for German, but ローベルト sounds closer; it has to do with word stress).