There are some unpleasant truths you learn very quickly when you get to the reading-real-things part of Japanese.

1. The Japanese consider spaces to be optional.
Corollary: People are lazy. If something is optional, it won't be done.

This is one of the few things that gets better when you learn kanji. Before that, it's one long string of kana where you just have to sort of insert breaks at random and test to see if it makes any damn sense. Once you start throwing kanji in, you can at least start by figuring that character + okurigana (kana used for grammatical endings) means it's a verb.

This also, incidentally, makes me uncertain why case markers in Japanese are taught as separate particles rather than inflections. There are short tags that come right after nouns to indicate nominative, accusative, genitive, etc. Left to their own devices, Japanese people will attach them to the end of the preceding noun. English-speakers will slap a space in between the two, to set the particle off as if it's a separate word. They're case-y enough that I used to mark up my German inflection tables in Japanese, because it's a hell of a lot easier to fit を into the little grid square than to try to microprint 'Accusative'.

Anyone out there start learning Japanese in a language other than English, and want to comment? I tried teaching Japanese to someone in French once, but there was a catastrophic mismatch of available verb tenses involved that made me glad I already had a grasp of the present progressive.

2. Japanese people don't finish their sentences any better than anyone else.

They're actually worse than most other languages I know. An honest Japanese-English dictionary would be half blank to account for all of the things that you say by pointedly not saying them. Do not let your attention wander during a conversation in Japanese or you will never pick it back up again.

3. Furigana lie.

Or, to be strictly accurate, furigana are not as definitive as you might hope. Furigana, or "flying characters", are the tiny kana that hover above (or beside, in vertical text) a kanji-ful word to give the phonetic reading. They're sometimes called Ruby characters, from the old name for the physical font size needed for them on printed pages. You see them most in stuff that's aimed at kids who have learned some but not all of the standard kanji -- my Detective Conan game, for instance, has a setting that lets me turn them on and off -- and sometimes on names, which have notoriously idiosyncratic pronunciations, even in works aimed at an older audience. You still occasionally see inline parenthetical pronunciations on platforms where the typesetter can't or is too lazy to dig around and rig it up properly, but XHTML supports Ruby markup now. I'd put it as no more convoluted than CSS, but I was at one point teaching a mini-class in CSS to a bunch of my schoolmates, so YMMV.

The thing about Ruby script is that you can make it say anything you want. Content isn't restricted to kana; really all it does is render a specific string of text at half-size and auto center/justify/kern it over a specified span. Chinese texts will sometimes use it for pinyin or bopomofo, and Korean ones will occasionally mark hanja in hangul. Wiseasses and people who write weird fiction will take unfair advantage of that. It is entirely possible to use Ruby markup to do this:


Where, if your browser responds properly to Unicode and Ruby, the kanji say "Super Magical Multi-Dimensional Blueprints of Heaven And Earth", and the furigana says it's to be read as e, which means "drawing".

Good luck with that there sci-fi manga, champ.

4. A lot of incorrect things become correct in context.

Most language classes are prescriptivist in nature, i.e., they give you a book that details how words should be used, and expect you to follow it or feel crushing guilt for being wrong. Then you get out into the real world and discover that sometimes people do things wrong for effect.

(I've had language teachers who were unable to distinguish between me being wrong by accident and me being a wiseass on purpose. Drives me bonkers, especially since the point at which I can do that is the point at which I start feeling competent. One of the reasons I liked the teacher I had at NAU so much was that she could generally tell when we were being screwball.)

Japanese, for example, has a lot of transitive/intransitive verb pairs. There's a set of matching bookend verbs for to open/to be opened, to close/to be closed, to move/to be moved, etc. They're approximately equivalent to active and passive voices in English. The thing moves, the thing is moved. You're told very sternly that it's incorrect to use the intransitive form with inanimate objects -- you can say ドアを開ける, doa wo akeru,  (someone) opens the door, but you cannot say ドアが開く, doa ga aku, door opens (itself).

Except if you read a lot of weird stuff, doors totally can do that. The second sentence is creepy as fuck, because doors aren't supposed to open themselves, but that thought does exist and can be uttered like that, and if you do indeed mean the door is opening itself, then it's not semantically or grammatically wrong.

They also tell you things like that the character ん only appears at the end of a syllable, and never appears by itself. Sure it can, if your speaker is making a small confused noise! It's the one English spells 'hm?' Only the interrogative version, though; the 'whaddayaknow' statement version is usually ふむ.

5. Kana spelling is not much better at being phonetic than any other kind.

Supposedly, modern Japanese spelling -- kanji aside -- was dragged back in line with modern Japanese pronunciation in about 1900, and then beaten completely into submission after WWII. To be fair, it's a lot better than it used to be. There were a lot of cases like verb endings that were spelled ーませう -maseu and pronounced ーましょう -mashou. To be critical, about half of the various grammatical particles are still written wrong (は, the subject particle wa, is written with a kana that normally says ha, even though わ normally serves for wa just fine; へ, a dative particle that's read e, is written with a kana that is normally used for he, even though there is a perfectly good え already; and を, the direct object marker, is read o now and used to say wo when that mora existed, and written with a kana that is no longer used for anything else anywhere in the language. This is all super fun when you're trying to enter Japanese phonetically in Windows using a QWERTY keyboard). There are also assorted other problems, like Japanese pretending that they have only five vowel sounds (they don't) which are always pronounced (they aren't) and that the language has no tonal accent (it does).

This isn't even getting into accents, dialects, and idiolects. Moggie and I were recently watching Magic Kaito 1412. The main character, Kuroba Kaito, suffers from chronic teenage-boy *mumblemumblewhatevernstuff* overlaid with one of the worst cases of sarcastic vowel drawl I've heard in ages. The subtitles give the misleading impression that he's speaking perfectly clearly; he's not. Japanese people are as prone as anyone else to taking a word like 面白い, omoshiroi, which really is supposed to be pronounced exactly as the spelling suggests, and mangling it into something which should properly be written おもしるぅぇぇぇぇ~ out of sheer bored snot-nosed sass. The subtitles render that "Sounds fun", but I'd bring it across almost exactly as "Eeeeeeeenterestingk," with the same rather loose association to what the word originally was.  Moggie apparently cannot understand a single godblessit thing he says.

Osaka-ben has some strange vowel changes involved and swallows a lot of less-pushy consonants, particularly near the end of words. No one ever notes this when it's written out. The pitch accent is also bizarrely different and varies over a couple of different axes that I have yet to figure out. Hattori Heiji, in Detective Conan, is from Osaka and voiced by a guy who is also from Osaka, and Moggie apparently can't understand him either.

6. Fuck names.

Just fuck them, All of them. Even native Japanese people can only take blind stabs at figuring out how name kanji are supposed to be pronounced. There's a Detective Conan case where some dude manages to elude capture for years by putting on a dress and changing how his name kanji were read. Seriously. There's a reason that my Japanese monogram is a single character that's read the same as a a male name, a female name, a color descriptor, and a goddamn noun,.