Is it okay to forget mnemonics?


I remember reading that mnemorics are only supposed to help you remember something in the short term and I just started actively noticing that fact.

For example I’ve just had a vocabulary in combination with “correct” and I for the life of me couldnt really remember what I was thinking when remembering that kanji. All I remember is that it’s either pronounced shou or sei.

Is it okay to forget what mnemoric you have with the kanji, as long as you remember the sound, or will it bite you in the long run?


You said it yourself. The mnemonics eventually get replaced with the actual knowledge. They are a bridge toward you knowing it, especially when you have no basis for making sense of all these funky lines initially.

But eventually you do build that framework and can look at something like 議 and say “oh, that’s easy to remember, it’s just two things.”


Ty mister Leebo,

I can do my 100 reviews with an easy heart now


Does anyone do WaniKani without the mnemonics?
So far I think I value WaniKani more for the well curated kanji/vocab list and design and have not really read a single mnemonic.

I’ve not really seen many kanji that I don’t already recognise or have some familiarity with yet though.


exact same question i asked early on ! good to see other people going through the same :wink:

I think i was at about the same level as well.

it still worries me slightly when I know the readings but can’t remember the mnemonic :cold_sweat:


I stopped using most of them around level 30 or so, not because I wanted to (the mnemonics are incredibly helpful), but rather because the radical system doesn’t accurately reflect how I think of kanji anymore. For instance, I just did the lesson for 剰. Wanikani lists the radicals as pine, mouth, and ribs. However, when I look at it I see the kanji for ride 乗 plus the right-hand side radical variant of sword 刀. So the mnemonic about ribs, pines, and mouths doesn’t really help me and I have to come up with my own about riding and swords. I also noticed that both 剰 and 乗 have the onyomi じょう which means 乗 is probably acting as a phonetic component, so I can use that to remember the reading rather than a mnemonic.

I wish Wanikani didn’t slavishly adhere to the radical system because it would be much simpler if they also allowed kanji to act as radicals. They often add kanji back in as radicals which helps, but you still end up with some unwieldy mash-ups like 惨 which is listed as being fish stick, hair, big, and pile, but it’s much easier for me to remember it as just being left-side heart + participate 参 and I just build a mnemomic around that.

I recommend doing the mnemonics as much as possible for as long as possible though, because they really are helpful when they work as intended.


I did the same at first, blissfully ignoring mnemonic. I think that was because I was already familiar with ~200 kanji, so my accuracy was very good, which gave me false impression that just doing review was enough…

But, for me, this approach started to break down passed level > 10. Lot of totally new kanji start to show up, lot of big kanji with many components like 能 or 疑 too, lot of kanji very similar to the basics one but slightly different…I couldn’t keep it anymore, my accuracy dropped down, so I felt back to mnemonic. And it was surprisingly very efficient !

There is a few caveat, as listed by rhet, but they are still much more efficient that I would have imagine.


Yeah, I’ve definitely found this same problem already. I had 植 come up, and immediately thought of the kanji for ‘fix’, 直. But that hasn’t officially been added as a radical (at least, not for me yet), so it broke the right-hand side down into the component radicals once more. This makes sense in terms of writing I suppose, but I’m definitely starting to add my own mnemonics where kanji like this form one half of the new kanji.

I don’t know whether they should make more kanji into “WK radicals”, or just automatically start using kanji in mnemonics for later kanji, but I think something could probably be improved here!

Oh, also, I highly recommend the composition userscript, as it definitely helps to make those connections with the tone marks.


Totally agree, it’s the exact same caveat I have with WK mnemonic system.

Why not always building mnemonic out of subkanji when possible ? Mnemonic would be so much simpler and also I guess would reinforce the knowledge of previous kanji.

Why is WK totally ignoring phonetic component, the historical built-in mnemonic system at the heart of kanji (admittedly, very messy) ? 惨 is good example, according to jisho onyomi is san or zan wich is not surprising at all because onyomi of 参 is also san… Just feel so weird to not try to reuse this knowledge.

I wonder if the wanikani team explained their choice somewhere about those two points ? Maybe they see adherence to strict consistency with their radical system better for learning ? Maybe they deem phonetic component too unreliable to be useful ?


Once you start to naturally recognise the Kanji, then yes it’s pretty normal to forget the mnemonic. If I find myself struggling to remember an older one, I might either refresh my memory on the mnemonic or come up with a new one. Ones I’m really familiar with I just straight up read without having to think about it, which is the goal really.

Totally agree about not always using WK’s definition of the radicals. Especially later on there are quite a few Kanji that are basically other Kanji with something added on, but WK breaks them down a little too much. 剰 mentioned above is a good example, I immediately thought of ‘ride’ with a rib added to it, but WK overcomplicates it - maybe because 剰 isn’t one of their radicals.


I think that’s probably the answer to that part of your question. I use the script which gives you information on the tone marks, and honestly it’s just an occasional bonus. I don’t know enough about the subject, but I don’t think it’s “at the heart” - lots of kanji don’t seem to have been constructed in this way at all. And lots of the tone marks seem to only indicate the reading for some of the kanji they appear in.

I imagine it would just add a whole extra layer of complexity that would only occasionally help you out.


I use some of them, but to be honest most of them are so weak (not a slight on WK, they go to a lot of effort and I can’t think of anything better) that I can’t remember them. The ones with more visual, descriptive detail are easier to remember but some of them which are just characters saying a random English word that sounds slightly like the vocab/Kanji reading just will not stick in my brain. For items which I struggle to remember that don’t have a useful mnemonic, the fact that I always get them wrong becomes a mnemonic in itself!


I found a really cool blog post about this :

From what I understand from the blog post, it still fair to say construction with phonetic component was at the heart, because historically 90% of all hanzi (not just kanji) were created with phonetic component… But yes, in modern japanese, so much time have passed, so many evolution, divergence, reinterpretation, accident of the language that only 20% to 35% kanji have at least one onyomi that can be linked to a phonetic component. It’s much lower than I thought, and probably explain why WK don’t bother with them.


This was incredibly interesting, thank you!

Yes, I think that the imperfect components are also, as the post suggested, actually worse than having no information at all. It can lead you to mis-guess readings. So on top of a large majority of kanji not having much modern connection to the phonetic construction, I can see why it’s more hassle to bother with it than not.

It’s definitely a helpful concept to be aware of though! It’d be nice if they just gave you a hint in passing in the reading mnemonic for the perfect ones, like “you might notice this is the same reading as these other kanji, because they all include X”.