Hey, also aphantasic here! A lot of guides focus on a pictorial way of remembering characters. That obviously doesn’t work for you or me. Considering I’ve learned all of the Jouyou kanji by now, I am in a good position to recommend my approach and to chat about what works for me and what might work for you. I apologise for how long this reply got, but there’s a lot to unpack.
Before I get into it, cards on the table, I don’t like RTK as a method, mainly because of its frontloaded approach to memorisation. But its mnemonic system and the decompositional approach to kanji absolutely works better for you and me than basically any other approach out there. To recommend something similar that has less of RTK’s flaws, I’d suggest looking into Kodansha Kanji Learner’s Course by Andrew Scott Conning. Conning takes a more holistic approach in ensuring you learn the parts, the vocab and the pronunications when you learn the character, as opposed to Heisig’s learn the meanings first, and the sounds later approach in RTK which doesn’t even get as far as vocabulary. Conning also provides workbooks for practicing writing the characters, and graded readers for reading the characters in context with their vocab. It’s a good all-round approach if you’re willing to buy in, but it is not in Jouyou order, which can be a bit of an obstacle if you’re JLPT oriented (Then again, RTK is not in Jouyou order either)
Anyway, some good news, character recognition does not rely on your image memory. At least, not once you know a character. It instead relies on your language processing, same way you can recognise written words in English, despite the fact that they are ultimately images on the page. Unfortunately, getting to know a character takes repeated exposure as comprehensible input. So just cramming characters over and over again doesn’t necessarily help with that.
It’s worth knowing that kanji can be broken down into a few different types, each kind of demands its own approach, since they are composed differently. So here’s a rundown of each of the types:
First, 象形文字 or pictograms. These are the hard ones for people like us because they’re basically drawings of the thing they represent. Even at the N1 level, I’m still screwing up writing 馬 and 鳥 Thankfully the fewest number of characters are 象形文字, but unfortunately they are some of the most common ones. For these, you’re going to want to practice writing them until they’re part of your muscle memory. Since picturing it doesn’t work, we have to rely on the physical feel of writing it. Your recognition should go up the more you write these characters. You’ll find a lot of methods targeting beginners talk up the pictographic nature of Japanese, since the first few dozen you learn are like this. This has multiple backfires in that it discourages aphantasic people by repeatedly telling us that kanji are pictures, and by setting people up for failure when they move onto the overwhelming majority of characters that do not behave like pictures.
Second, we have 指事文字 or simple ideographs. These are more abstract simple characters, very common in the N5 and N4 range. Characters like 上下一二三, where the character is an abstract representation of the idea. Funnily, these are a little easier for us, since we can reason about them (三 is just three lines, 上 is a stick above a line, etc.) This is where the pictographic methods start to waver and where the bulk of the setup for later character types begins with other methods.
Third, 会意文字 or compound ideographs. This is where things get much easier for us and is where the majority of Kyouiku kanji are found. 会意文字 are composed of parts we’ve already learned, so this is where we can start thinking of characters less as pictures, and more as composition. The question stops being “what was the character for ‘eye’ again?” and becomes “what is the character for ‘rest’ composed of?” It’s person 人 and tree 木, so 休. Heisig and Conning’s methods absolutely excel in this block of characters.
Lastly, 形声文字 or Semasio-Phonetic characters. These are the bulk of characters toward the end of the jouyou list, and where pictographic learners really struggle, whereas we can really benefit. Right now this seems like the set of characters you’re struggling with most, characters that differ by their radical but are otherwise identical are majority this type. 形声文字 are composed usually of two parts, a radical indicating the “class” of word, and a component indicating the pronunciation of the character (the on’yomi, specifically). With these characters, reasoning becomes “it’s an animal, so beast radical 犭… and it’s pronounced ロウ (艮) so it’s 狼”. Heisig and Conning both work quite well here, though Conning has admitted that Natalie Hamilton’s “Kanji Code” does a much better job of making sense of these characters and wishes he’d incorporated those ideas more in his work.
So with all of that boiler plate, what’s my approach exactly?
First, build up a mental character inventory. Learn radicals and pictographs the same way I learned my ABCs as a child. Practicing writing them until they’re part of my brain. There’s more radicals and pictographs than there are letters in the alphabet, yes, but putting in the grind here makes later kanji much more intuitive to learn. Plus you don’t have to do all of this up front. Pick kanji for the level you’re studying for, and learn those radicals and pictographs. N5 has maybe 100 characters, about a third of which go on to become part of the latter two thirds. N4 has maybe 300 characters total (incl. the N5 ones), and the bulk of which are composed of the N5 characters and some of the earlier N4 ones.
Second, learn the way characters are composed. Generally characters are composed in a few well-defined ways, such as in columns, rows, one portion surrounding or wrapping around another, and so on. Japanese has specific terms for these, like hen and tsukuri, but just being able to recognise and break down a character into its parts and their positions is enough.
Third, start thinking of characters as characters composed of smaller characters, not pictures. You’ve already got a taste of this with Heisig, but in general, when you see a character, don’t try and see it as a pictographic whole, see it as “thread 糸 to the left, right is wings ヨ over water 水 which is pronounced ロク. That’s midori 緑” As you grow more familiar with a character you’ll start to need the composition less and less as the whole character itself becomes part of your brain’s language processing. It becomes “left 糸, right ロク so midori 緑” up as far as “ah, 緑” with no conversion step. Mnemonics are a stepping stone to get to this simpler and simpler recognition.
Lastly, kanji are not the end goal, vocabulary is. Kanji recognition actually goes up when you focus on them in context. It’s a lot harder to remember kanji in isolation. What is the difference between this kanji, and this other one with the same radical, similar meaning, but different reading? Hard to say. Much, much easier to be able to say “this is the kanji used in X word, that kanji is used in Y.” or “These two are used together to make one actual vocabulary word.”
