Thoughts about Wanikani

Wanikani has been with me since the very first month of my Japanese-learning journey, and I was a big fan for a long time. Over the years my opinion of it has gradually changed, and at this point I feel like I can’t recommend it in good conscience without some heavy disclaimers. I don’t think it’s a bad platform by any means, and there’s a lot to like — I finished the whole course and I’m quite certain it helped me avoid some common pain points — I just feel like if I’m going to say “wanikani worked for me, it’s pretty good”, it’s only right that I be able to point to a list of criticisms to temper that recommendation.

I’ll also be laying out the strategy and schedule I used to get through the second half of the course quickly and efficiently while also working full time.

“You can’t ‘learn’ kanji!”

Let’s get this out of the way. People who hold this view are insufferable. They tend to define “learning kanji” in some absurd way like “knowing all the readings, the stroke order, the key meanings, and the exceptional meanings in irregular words”. It’s almost as if they were going out of their way to give it the most unachievable definition possible. Hmm…

Other people will try to get you to admit that “learning kanji” is actually “learning words written with kanji” or that “a kanji’s meaning is its kunyomi”. This is a trap. While I do agree with learning kanji readings via vocabulary (more on that below), it is undeniable that native Japanese speakers “know kanji” individually because they are capable of understanding the meanings and pronunciation of completely made-up words formed entirely of kanji, like 魔哭鳴斬剣1from 幽遊白書 or 絶刀空閃2from 鍵人. In the majority of cases, these words are made up of on’yomi readings. If you can’t recognize a kanji when it’s isolated from the words you learned it in, you won’t be able to read these words.

Not all kanji can be used to write standalone words. Not all kanji have kun’yomi readings. If you want to approach native-level literacy (or even just enjoy shounen manga without skipping over all the names), you will have to “learn kanji” at some point.

Personally when I say I “know” a kanji, I think of it like how you can “know” a person. You can identify him by his face, but maybe if he changes his hairstyle or takes off his glasses, you might not recognize him at first. You might know where he works and a couple of his hobbies, but you might still be learning surprising new information years into your relationship.

You do not need to know everything there is to know in order to say that you know something.

Radicals (good)

Wanikani teaches kanji by breaking them down into radicals3Technically they should be called components, but this is really a very minor distinction in an age of digital dictionaries where identifying the literal “radical” of a kanji is a useless skill.. Some people don’t like this, but I would never recommend any other way. I’ve seen people argue for just learning kanji by their general shape, and then turn around and say that they thought 待つ and 持つ were the same word with two different readings, or that 少し and 小さい shared a character.

Breaking down kanji into components so you can know exactly where to look to distinguish between two visually-similar characters is extremely helpful. I’ve helped many people unlock a significant boost to their kanji recognition ability by giving extremely simple tips like “the left-hand component of 持 is a variant of the hand character”. Ignoring composition and focusing entirely on general shape may work for some people (though I don’t believe it works as well as they think it does), but it doesn’t work for me and I know I’m not alone.

If I have one gripe about Wanikani’s radical approach in particular, it’s that there are far too many of them. There are only 214 orthodox kanji radicals, but Wanikani introduces many of their own or will teach previously-learned kanji again as a separate radical lesson instead of adjusting the curriculum to mention that “sometimes kanji can appear as radicals”.

Mnemonics (pretty bad)

Mnemonics are another divisive topic. Personally I don’t mind them. I find that at worst, they don’t help at all, and at best, they help initially and then eventually fade from my memory once they’ve done their job.

Wanikani’s mnemonics are arguably the majority of the product, and unfortunately they’re not very good. They start off strong (maybe a little too strong sometimes) but as the levels go on it feels more and more like they’re being phoned in. My go-to example will always be for the character 備: “The leader put flowers on a cliff. That is their task. It is to provide and equip the cliff with flowers.” This is completely useless. The division between story and meaning isn’t obvious at all, and the story itself has no relationship to the meaning.

At a certain point, once my “kanji sense” had been developed a bit, I stopped paying attention to the mnemonics at all. The course still worked for me. The problem at that point is…what exactly am I paying for?

Price (inexcusable)

I started Wanikani around the same time I got my first full-time job. Suddenly I was making more money than I knew what to do with; I had no student loans, no car payment, no utility bills. Dropping $300 on a lifetime subscription was an easy decision for me at the time, but for many people, that price is literally unthinkable. Even at $9/mo or $100/yr, it’s still pretty steep.

When Wanikani first started back in 2011, it might have been worth the price, but nowadays with so many other good options out there for a fraction of the price, Wanikani is basically impossible to recommend to anyone who doesn’t have a lot of disposable income.

Pace (good)

People often complain about Wanikani’s pacing, and I always tell them “finish level three before making judgements”. With over eight thousand lessons, you would have to follow a consistent pace of 20 items per day in order to finish in a year. While this might be slow for ankibrains who want to grind through RTK in a single month, it’s pretty rigorous for the rest of us — not to mention the fact that trying to play the numbers game to “get done with kanji” is the number one cause of burnout based on my observations.

Readings (great)

Wanikani teaches you one “main” reading per character, and uses vocabulary to both reinforce that reading and to introduce other readings in context. Typically the way it works is:

  • If the kanji is most commonly used in on’yomi compounds, the main reading will be on’yomi
  • If the kanji has a lot of irregular kun’yomi readings, the main reading will also be on’yomi
  • If the kanji is commonly used to write a standalone word, the main reading will be the kunyomi of that word

This seems to be something a lot of people don’t understand, so let me be completely clear: Wanikani is a “learn words” approach. Criticism is all well and good but it must be accurate and honest criticism.

Meanings (great)

Kanji mean things. Sorry! I know this is hard for some people to accept. But they do. Sometimes they can mean multiple things (like how 着 can mean both “wear” and “arrive”). Sometimes the meaning is only ever a component of other words (like how 然 means “as though” and cannot exist on its own4Not including archaic speech/ateji, which is a whole other can of worms.).

Trying to teach kanji as simply “letters that can be used to spell words” is short-sighted and does not reflect the way that native Japanese speakers think about them. The meanings Wanikani assigns to kanji are helpful and pragmatic. None of that “decameron” nonsense, or teaching 脚 as “foot” and 足 as “leg”5The exact opposite of what it should be! like Heisig’s RTK.

Coverage (good)

Wanikani doesn’t cover all the jouyou kanji, but it does get pretty close. Honestly, I would prefer they taught a little bit less. 梓 is simply not a good character to learn up-front. But 90% coverage is darn good, especially considering that every character has an average of three words associated with it. Six thousand vocabulary words and two thousand distinct kanji is highly respectable.

Recognition, not recall (good)

Handwriting is a waste of time. Recalling kanji from keywords (which must be unique and are therefore not very useful) is not a useful skill. It’s commendable that Wanikani requires neither.

Wanikani teaches you to recognize kanji and quizzes you on their meaning and reading, both of which (as demonstrated above) are practical and useful things to know.

SRS (good)

Wanikani’s SRS intervals are different than Anki and other similar solutions, but I like them. They seem to be deliberately designed to support high saturation early on, as well as a high level of predictability. Here’s the schedule I used to use when I was trying to be as efficient as possible in order to grind through levels quickly:

  • Do my lessons for the day at 9am, along with any reviews that are due.
  • Do my next batch of reviews at 1pm (SRS interval 1, four hours later). My new items from this morning are all mixed in here.
  • Do my third batch of reviews at 9pm (interval 2, eight hours later). This batch of reviews includes not only my lessons from this morning, but also (assuming I got them right each time) my lessons from the previous morning’s lessons.
  • At least once per day, I would get my due reviews down to zero.

By doing my lessons at a fixed time, and my reviews at three set times throughout the day, it’s possible to get maximum short-term exposure to new material, increasing the chances that it’ll sink in quickly. It also allows me to plan ahead and not feel like I need to be constantly checking for new reviews.


Do I hate Wanikani? No…but I don’t exactly love it anymore either. If I had to do it all over again, I would definitely use something else. It doesn’t really sit right with me to spend literally 10x the price of KKLC for essentially the same methodology, but with mnemonics I won’t use. If the plug-and-play nature of the course appeals to you, and you can afford the price, and you like the first three levels (free to try), I wouldn’t advise against it, but that’s about as much as I can say.