I think I’d even go further and wish that we could skip the kun and on reading cards for just their meanings and let the vocabulary itself guide the readings. Where over time the vocabulary and maybe breaking them down into components will naturally reveal the patterns over time.
3 Likes
Hi all 
Just wanted to let everyone know we’ve been watching this thread and taking notes, and I’m really sorry for the disruption the reading updates have caused.
First, please know that we’re not planning to change tons of readings to kun’yomi. We’re still following the on’yomi-first approach when it makes sense, but we also want to be a little more flexible moving forward. But the feedback here has definitely helped us to rethink where that threshold should be, in terms of justifying a reading change to kun’yomi, so thank you 
So why teach kun’yomi first at all? Without going into too much detail (it’d be a wall of text), there were several factors we thought about and debated extensively before the 明 and 夜 changes. These were related to fixing some existing issues, like radical and kanji scaffolding (we wanted them closer together) and multiple readings in the same level (we wanted different readings taught further apart). The biggest problem was that of on’yomi vocab available for reinforcement — all of it was either formal, not super common, or advanced for the early levels 明 and 夜 are in.
We know the system could better support these kinds of content updates, and we’re trying to figure out features or changes to make them less disruptive. Some of you mentioned allowing learners to reset an item and start fresh, or linking pop-up messages directly to the updated page. These are great suggestions we’ve taken note of.
I totally understand that inconsistency can be confusing and discouraging, and ideally, everything would follow a perfectly consistent system. But kanji and vocabulary come with a lot of exceptions, and trying to fit them all into a single system doesn’t always lead to the best learning experience. We’ll keep looking for ways to balance structure and flexibility, and to present these changes in the clearest possible way for all of you.
25 Likes
Thank you for your extensive reply! It’s good to hear you’re following the discussion.
Personally for me, it also disrupts my learning if a reading I have learned before is suddenly changed, including the mnemonic. Would it be possible to keep the same reading for the people who have already started learning the kanji with the onyomi? Often when I forget a reading, I will check the mnemonic, but in the case of 明 and 夜 the old mnemonic is no longer there. It doesn’t help my learning to have the tools I am using for learning changed halfway through. Another option would be to keep the onyomi as a correct reading as well, since that’s technically a correct reading for a kanji, no?
Thank you for your consideration!
4 Likes
I understand where you’re coming from but I also think that you may be overthinking this and trying to optimize for the wrong thing. In this case the problem seems to be that you decided to introduce 明 very early on level 3, but then you lack kanji like 説 (level 17) to reinforce the onyomi. My opinion is that, in this case, maybe you shouldn’t introduce 明 that early.
Would it be annoying for learners to have to wait until level 17 or so to learn such a common kanji? Well, yeah, in the same way that it’s annoying to have to wait until level 16 to learn 書く, level 25 for 状況 and level 30 for 遊ぶ。
There’s no way around it, basically every single one of the top 1000 kanji by frequency is extremely useful and should be memorized as early as possible, but most people won’t be able to do that comfortably in less than a year. Forcing some of those kanji super early at the cost of teaching weird readings is in my opinion not worth it because that can cause lasting damage to your ability to guess readings correctly. I mean, let’s face it, it’s not like you will be able to read basically anything at level 3 anyway, 明 or not, so you might as well postpone that for a few months if it means that you’ll learn the kanji “properly”.
That’s true of course but some of the issues here are self-inflicted. Like deciding to teach meaningless truncated readings like 明（あ） or 取（と） or 涼（すず）. I have already expressed that multiple time in this thread (and other threads) but I’ll just once again insist that it is in my opinion a complete waste of time to learn these readings this way, and the only reason it’s done like this is because of arbitrary technical limitations on WaniKani’s side. If you really think that it makes sense to teach the kun first then do it, but teach the actual kun, not these meaningless sounds.
Actually I almost regret having mentioned 遊 above because it seems like a prime candidate for the kunification process… If I see a content update soon where its main reading switches to あそ I’ll blame myself.
16 Likes
The problem is that the memonic for the めい reading of 明 is just gone. I look at the memonics for higher level words with 明 and it just says “It is a jukugo word so onyomi readings”. But I didn’t learn the onyomi reading. At least put the memonic for めい in the lowest level word its used in. めい is too common to not have its own memonic.
12 Likes
The entire point of wanikani is to teach kanji in the order of complexity of design
3 Likes
Sticking to just the onyomi would be so much easier for me, and if there isn’t a jukugo word around that level then I feel like that’s just how it may be sometimes. This inconsistent butchered kunyomi approach is really confusing. And to some extent how the kanji are distributed throughout the levels is arbitrary. So at this point it just sounds like you all have dug this hole and you want to keep on digging it. Which… that’s fine. It’s your all’s tool to develop. With my other issues with wanikani though, I just really wish there was a better tool out there other than just doing a ton of work setting up and using anki to fit my needs.
7 Likes
Kinda, but also not really: https://www.wanikani.com/kanji/又
It’s part of it of course, but relative frequency is clearly a big factor too.
7 Likes
Mmm, if you want to see an “increasing complexity of design” order, look at the RTK order, which makes no concessions at all to frequency of use, and happily teaches uncommon kanji like 旦 (wk level 55, rtk kanji 30) early because it’s simple and very common kanji like 馬 (WK level 8, rtk kanji 2132) late because it didn’t fit into the “teach a component and then all the kanji you can now make with that component” order any earlier.
4 Likes
I think I’d want to see it ordered by frequency even if it’s a bit challenging to see very complex items.
1 Like
Also thank you for the reply. To emphasise what someone else said, maybe the solution in this case would be to teach later when there is more supporting vocab?
WK and tofugu.com are very clear that WK alone won’t teach you Japanese, so most learners are going to be using other resources as well, which will cover higher frequency vocab earlier.
I very much value the principle of teaching what most enables you to guess accurately. Ie teach the most common on reading. Ultimately that gets you the widest coverage the fastest.
The scaffolding approach is also great, eg teaching entrance/exit/doorway in close succession so you can see and reinforce the pattern, but I can’t believe this gives an ordering by vocab frequency. WK users already delay learning loads of common words until they build on what has already taught, so I don’t see why these frequent words can’t wait until they’re taught in conjunction with the scaffolding.
1 Like
So just reviewed 夜。 Put in や。Wanikani gives a message saying something like Yes that’s a reading, but what is the most common reading. It’s not as bad as some of the truncated readings like tsu for 付く。
I had just noticed this morning that that particular input (や as a warning reading for 夜) is a fresh intentional design choice that they literally just implemented today: Thursday November 13 2025 Content Updates
1 Like
A solution (?) is to accept On’yomi anyway (rather than warn) when a Kun’yomi is taught, but not vice versa. In contrast, if On’yomi is taught already, don’t accept Kun, warn only.
Though, imo, it might not be fair to just accept all On’yomi, when there are multiple readings. Warn can be used well.
Isn’t in that recent update, some Vocab are being misclassified as Kanji?
6 Likes
Just reviewed 北。I put in ほく。The message was something like we just changed the reading. Please check the page for the new Kanji reading. Fun. Fun. Fun.
3 Likes
Yes that was another big set of content updates they made over the last month or so for the cardinal directions.
They don’t even accept ほく at all now as a reading for the 北 kanji? I guess I’m fortunate to have had these burned by the time this set of changes went out.
2 Likes
Thank You!!! I need to let this go, but you have said exactly what I believe–that users who got through the program before the changes will have a huge advantage compared to new users in their kanji learning journey with wanikani. I know wanikani will not change. 
2 Likes
I mean, very marginally so if you think these new changes are extremely detrimental, but really we’re talking here about a few kanji out of over 2000.
1 Like
There are more every week. People. who have a lot of burned kanji won’t notice it.
3 Likes
Why not accept all (any) reading?
After all, of there is one thing widely agreed, is that the real learning is through real words (vocab).
While there are pros and cons for prominently teaching one reading over another at the kanji tile level, why not accept all possible readings in quizzes? Both よる and や are proof that 夜 is recognized, no?
(What would be nice also, is to use real vocab readings for kun, et, not just “a" for 明, but “a.keru” or “a.ke” (or “aka.rui”, or “mei”, etc. all should be accepted imho. But clearly not “a" which is not a word)
1 Like