I joined Wanikani on the premise of teaching kanji in order of visual complexity rather than meaning complexity. However, I started to notice some oddities in the way kanji are grouped by levels.
Let me give an example:
Radical: 士 (samurai) - level 4
Kanji: 士 (samurai - し) - level 13 (?)
Kanji: 仕 (doing - し) - level 4
This is highly problematic because 士 is a phonetic component for 仕 and a bunch of other kanji, but because they are not taught in sequential order Wanikani tries to teach me the same reading し through multiple mnemonics which is rather inefficient. Reading aside, 仕 is also more visually complex than 士, so I don’t understand why they are so far apart.
I am trying to assess how often this happens; if it’s a common occurrence I will be a deal breaker for me and I’ll have to look for another learning source. Did you notice it frequently?
I don’t think it happens very often, but it sometimes happens. Be aware that while WaniKani was designed with the intention of teaching kanji in the order of complexity, some misses might be there. The samurai kanji likely being one of them.
@Mods I think the arguments are sound here. @saller123 would it make sense to move the samurai kanji to level 4 in your opinion?
It’s not perfectly based on visual complexity, partially because that is a subjective measure and partially because it would be incredibly useless. WK mostly folows complexity order, but it also tries to prioritize words or kanji that are more common. You will be able to feel this effect over time as well, because later levels will start to feel more and more abstract in their vocabulary.
Also, WK doesn’t use semantic-phonetic composition at all, each kanji with that reading will in fact get a completely new mnemonic based on its radicals and kanji rarely lean on each other for memorization help. However at the very least they will all use sheep in their reading mnemonics. (at least afaik)
as @Gorbit99 says, WK doesn’t rely on semantic-phonetic composition at all, that’s not how they teach kanji. However, there is a very cool user script that supplements WK lessons with that information and I highly recommend installing it:
I don’t think it would make much sense considering how many kanji can have the same reading due to semantic-composition. That’s simply not how WK is organized at all.
My take on it, is that this is because “samurai” isn’t that useful compared to learning “doing”. This happens all the time on WK. It’s a matter of priority.
If you move a kanji forward, then you would also have to move at least two pieces of vocabulary forward to help remember the reading. But these will probably also contain other kanji, which then might have to be moved forward, and you quickly get a chain reaction.
the order in which WK teaches items is all over the place. it loosely follows frequency order, but only loosely. it loosely follows jlpt and joyo order, but only loosely. by building kanji out of “radicals”, it approximates ordering by visual complexity, with some big outliers…
so if you want to learn your kanji following some strict progression, you won’t find that at WK. on the other hand any criterion you pick to order your kanji by will have things which make no sense if you examine the order by some other criterium. i.e. if you order by frequency or jlpt, you have some visually complex kanji early on. if you order by number of strokes, you have some relatively rare kanji very early on.
WK at least uses a system. though it’s far from perfect, it’s decently good. at least for now, as long as they don’t break it by adding thousands of non-kanji vocab…
Hey linchou! Thanks for the ping! I think what you and ekg have gotten at is a pretty good explanation. I also think it might have a bit to do with the placement of the 士 kanji’s related vocabulary, which start at level 13, and you get another one every 5 to 7 levels after that. Moving the kanji to level 4 would probably require some vocabulary movement as well.
The radical is used in three kanji before level 13 which seems to justify its placement at level 4.