Japanese
At the end of 2023, I decided to start studying Japanese. It's been over a year, now, so I wanted to collect some of my thoughts.
Study plan
General approach
I watched Tokini Andy's YouTube video "How I Would Learn Japanese (If I Could Start Over)" at the suggestion of a friend and went through a bunch of reddit threads to figure out how to start.
Of course, the first thing I did was learn to recognize the kana. I used the Tofugu Learn Hiragana and Learn Katakana pages for this. They teach you to recognize a picture in each character and drill those mnemonics. I was worried that I would continue to think about these mnemonics constantly, but after reading the characters sufficiently many times in context, they start fading. They serve their purpose of getting you to remember the readings, then they get out of the way. I also used DJT Kana to practice for a few hours at the beginning.
For vocabulary, Anki is king. I decided to go with the "Tango" decks instead of the "Core" decks. The Tango decks (made by Nukemarine) are based on the Tango books. The books are basically lists of terms, their English gloss, a sentence using the term, and the English gloss of the sentence. Sometimes you'll get a related term or two (synonym, antonym, or just similar term) – these don't get their own sentences. One cool feature of these sentences is that they are "i+1", meaning that if you go through the sentences in order, you'll only encounter at most one new term in each sentence. Stephen Krashen's input hypothesis claims this is useful. Anecdotally, I agree.
The Anki deck consists of two types of notes: with sentences and without. The notes with sentences get two card types: "reading", in which you read the term and sentence and try to recall the meaning, and "listening", in which you listen to an L1 speaker read the term and sentence and try to recall the meaning. The notes without sentences correspond to those "related" terms. Since Japanese has so many homophones, they only have "reading" cards so you can use the kanji to disambiguate.
For grammar, I'm using the Genki textbooks. and Tokini Andy has videos that accompany each lesson for both Genki I and Genki II. I think that Genki is fine, but not superb. Common alternatives include Minna no Nihongo (みんなの日本語), Tobira Beginning Japanese, and Tae Kim's Guide to Learning Japanese. My typical flow has been: read through the lesson, take some notes, do some of the exercises at the end, and watch the corresponding Tokini Andy video to reinforce.
Kanji
For several months at the start, I used the Kodansha Kanji Learner's Course (KKLC) to study kanji. I was using this basically like one might use Heisig's Remembering the Kanji: learn a mnemonic for the general meaning of each kanji and focus on both recognizing the meaning for the character and recall the character based on the meaning. Each day, I went through the four characters on a page. I did my best to internalize the meaning and understand the connection to the mnemonic, looked at the readings, and read over the example vocabulary. I used this Anki deck to drill the recognition and recall (including proper stroke order). Initially I tried including a couple vocab terms for each character, too, but this proved to be extremely frustrating. This should have been a hint that, at the very least, the way I was using the KKLC wasn't going to work out.
Gradually I became disillusioned with the concept of learning meanings alone, totally disconnected from readings and vocabulary. When I came across new vocabulary with a kanji whose general meaning I had learned, I could sort of understand what the term was getting at, but I had no idea how to read and pronounce it. Using a tool like RTK or KKLC (the way I was using it, anyway, which is not how Conning recommends you use the course) is a fine way to gain familiarity with how the characters are constructed and how that construction relates to their meaning. I don't think there's any debate about that. However, by the time you've reached the end, you haven't necessarily learned much Japanese. I tried using the KLC Graded Reading Sets to supplement my study as Conning suggests. I believe this could have worked, had I stuck with it, but for me it was far too dry.
I had no problem finding others expressing the same frustration online. After 572 kanji and about 6 months, I decided that – for my learning goals – dedicated kanji study is not required. Currently I'm focusing on the two input skills: listening and reading. Eventually I'll add output (speaking and writing), but I doubt I'll add much kanji study in even when that time comes. Practicing the skill directly and consistently is what really helps. I'm encountering all of these kanji as I learn new vocabulary, and I reinforce them with Anki and reading. The intermediate step of memorizing a mnemonic for every character doesn't seem to be necessary, nor does it seem to help very much.
Two important caveats:
- Going through several hundred characters did introduce me to many of the components (radicals) used to construct kanji. It's hard for me to say how much this has helped and whether the time studying them would have been better spent elsewhere.
- If I were interested in physically writing (rather than typing) characters, I would of course need to practice writing the characters out. This isn't quite "kanji study" as much as it is practicing the skill of writing, though, so I think I would approach this by focusing on the content of what I'm writing and let the kanji serve as a means to an end. Writing them again and again as I create sentences would serve as a natural frequency-based spaced-repetition system in the same way that reading actual books does.
Listening
After around 9 months of study, I started listening to beginner-oriented podcasts during my commute to and from work. I found Japanese with Shun to be the most approachable. I listen to most episodes three times, and occasionally I go back and read the transcript later on. With each listen, I pick up more than the last, although I'm still not close to 100% comprehension. Progress is slow but quite real. I first tried adding this kind of listening practice into my routine after 4 or 5 months of study. Listening to 5 or 10 minutes of audio and picking out only a few words was discouraging, at the time, but I wish I had continued anyway. Yes, most of it flew right by me, but every time my brain recognized a word ("understood a message" in Input Hypothesis terms), some neural pathway got a little stronger and I advanced one minuscule click on my learning journey. Now I have that same occasional feeling with full sentences. Eventually I'll move on to harder content so I can continue to get these small "click" moments where my understanding improves.
A common recommendation is to watch Peppa Pig or other very basic children's shows. I have spent some time doing this, too, but I don't find it especially engaging. Still, it's useful: I remember hearing 自転車 in context while seeing a bicycle on screen after a couple months of study and feeling that word lock into place in my brain. That's how the immersion approach works, provided you can deal with the huge amounts of speech that you can't understand in between those moments.
Reading
The free Tadoku graded readers were a great way to start practicing reading. I went through most of the S and L0 books, as well as some of the L1. They're very short and use extremely basic grammar and vocab. The primary goal with these is to get extensive reading practice – reading a lot of material that's roughly at your current level of understanding so that you can get as much input as possible. This contrasts with intensive reading, in which you focus on understanding every last bit of the text. Extensive reading tends to be more enjoyable and feel less like studying, but both have their benefits.
After going through the lower levels of the free Tadoku books, I bought a subscription to Satori Reader. It's a website and app with a ton of stories, each split up into short "episodes". There are high-quality narrations from L1 speakers, in-depth grammar and usage explanations, and per-word translations that show the meaning as used in the context of the story. You can adjust furigana based on your current knowledge by either connecting to services like WaniKani or simply pasting in a list of kanji you consider "known". There's also a built-in SRS, but I haven't used this much since I already spend a lot of time reviewing cards in Anki. Going through an episode or two every day has been a great way to reinforce vocab, grammar, and kanji knowledge. You can also play the episodes in a sort of podcast-like mode where the app just plays through them back-to-back. I am experimenting with using this for listening practice, too, by playing episodes I've already read through once. I don't think I could have started this app much earlier than I did (that is, around the time I finished the Tango N4 deck) without being frustrated by the amount of words I don't know. Now that I'm at a level where I can use this, I am finding it indispensable.
Progress
- December 2023
- Learn kana; start Tango N5 deck; start Genki I; start Kodansha Kanji Learner's Course.
- April 2024
- Complete Tango N5 deck (vocabulary is at ~1000 words); start Tango N4.
- June 2024
- Stop Kodansha Kanji Learner's Course after 572 kanji.
- September 2024
- Complete Genki I; start Genki II; start listening to podcasts while commuting.
- November 2024
- Complete Tango N4 deck (vocabulary is at ~2500 words); start Tango N3; start using Satori Reader.