how can you understand japanese without subtitles?
four habits of spoken japanese stand between most learners and an unsubtitled scene, and each can be learned by ear. the japanese of films, dramas and songs is mostly not the japanese textbooks start with: it is casual, it leaves out who is talking about whom, it saves its verb for the end of the sentence, and some of it, especially in anime, is a stylised speech nobody uses at home. knowing them changes what you listen for, and the rest is familiarity: the same scenes and songs heard many times, with their meaning already known, until the sound carries the meaning on its own.
the japanese the textbook starts with
most textbooks for learners begin with the polite style, the verbs that end in ます and the sentences that end in です. japanese children learn it the other way round, informal speech first and the polite style later. a standard reference grammar says both things on the same page.
on screen, between friends, partners and family, the polite endings mostly fall away. what is left is shorter, and much of it is built from forms the first chapters of a textbook do not yet teach. when a line seems too quick to catch, it is often because it is in a style you have read about and rarely heard.
a sentence with no "you" in it
take a line from driftspeak's japanese volume built around mariya takeuchi's variety, from its track for one night stand: 街を旅してるんだね ("you're travelling around, town to town, huh."). the english has a "you" in it. the japanese does not. japanese lets the subject go unsaid whenever the situation makes it clear, and the situation usually does. a subtitle has to choose a pronoun; the speaker never did. listening without subtitles means letting the scene do the work the missing word would have done.
the same line shows the casual style at work. 旅してる is a spoken contraction of the fuller form a textbook would print, the kind of shortening japanese speech makes in a regular way. and the line ends on ね, a particle that softens a statement and asks the listener to agree, close to the "huh" in the gloss.
the verb waits until the end
japanese puts its verb at the end of the sentence, after the subject and the object, an order linguists write as subject, object, verb. the endings that make a verb past, negative or polite sit on the verb, so they arrive at the end too, and the particles that turn a sentence into a question or soften it come after that.
an english listener is used to learning early in a sentence what is being done and whether it happened. in japanese you often have to hold the whole sentence before it settles into a yes, a no or a question. this is part of why a line can seem to slip past: its meaning lands on its last few syllables, and by then the next line may already have begun. listening for endings, rather than for every word, is the habit that helps most here.
the voices fiction gives its characters
if most of your japanese comes from anime, some of it is not quite real. in 2003 the linguist satoshi kinsui gave this a name, yakuwarigo, "role language": the set ways of speaking that fiction uses to mark a character as an elderly doctor, a samurai, a noblewoman or a boy from the countryside. you recognise the type at once, and you would rarely hear that speech from a real person of that kind. it is worth enjoying for what it is, and worth not copying into a conversation with the person next to you on a train.
the useful part is that role language is consistent. once you notice that a character always ends sentences one way, you have learned something about how endings carry who is speaking, which is the same skill ordinary speech asks of you.
start with a scene you already know
the most useful material is a scene you have already watched with subtitles, so the story is no longer the question. watch it again without them and listen for three things only: where each sentence ends, whether the endings are polite or bare, and whether you can tell who is being talked about from the situation alone. do the same with a song you love, where the words already sit in your memory. stay with the same few minutes until they hold no surprises, then add the next scene.
this has a limit worth naming. a scene you know by heart is easier partly because you know it. the real test is a new scene in the same kind of speech, with the subtitles off from the start. if you can follow the gist of that, the familiarity has started to carry over.
city pop, said rather than sung
driftspeak is subtitles for your ears. vol. 01 is built around variety, mariya takeuchi's 1984 record, eight tracks long. each track is made of short japanese sentences from the record's world, each said in japanese and then in english. not one of its 333 lines ends in です or ます: it stays in the casual japanese of the songs, the japanese the textbook leaves for later.
where is the "you" in 街を旅してるんだね? listen to a sample of vol. 01, where the first three tracks have free samples, and go on to the third, whose sample opens on that line and ends on a longer version of it.