Mentalists & Memory Palaces: A better way to memorize Kanji

Mentalists & Memory Palaces: A better way to memorize Kanji

WaniKani is an excellent program. For years I have unsuccessfully memorized kanji and now I go through level after level with a quickness and a high rate of accuracy. Overall I have an exceptional memory, but when you get to the higher levels of the program, this becomes a handicap more than a skill. It’s in the little details that distinguish one kanji from another and if you are relying on the same pronounciation and same visualization that applies to eight plus kanji, things can get pretty tricky. Mnemonics is an excellent tool, but once you pass that 1,000 Kanji threshold, Mrs. Chou ちょうis more confusing than she is helpful.

Enter Memory Palaces.

I was searching for a way to make studying more efficient, forcing me to utilize all radicals of each Kanji each time in a very unique way. I wanted to also create strong associations across levels of similiar concepts or information that may be used in the same conversation.

Yan-Jaa Wintersoul. Maybe you have heard of her? She is a mentalist and Memory World Champion who despite not knowing me from anyone else who is a fan of her work, I consider her to be a personal best friend and mentor because of all she’s taught me.

“Moonwalk with Einstein” by Joshua Foer is a book all of you should read if you are serious about any form of memorization. You will quickly learn that memorizing a deck of cards is not much different as learning a Level of Kanji with WaniKani. Each card can easily be represented by a Kanji while each aspect of the card (color, face) is much like the radicals we learn. And by choosing a Memory Palace for each Level (Yes, 60 Memory Palaces) you can easily be transported to the proper Memory Palace and place within the Memory Palace to answer that kanji and all its respective associated vocabulary. We just need to store that information long term instead of short term as Yan-Jaa or any other Mentalist does with those competitions.

The reason Memory Palaces are so successful is because it ties visual symbolic representations of places in space and time, just as the brain normally stores information. Just as you are unaware that you are learning about the Fibonacci sequence from the famous mathmetician Lewis Carrol when Alice from Wonderland chases the white rabbit and falls down the rabbit hole, you can meet an angry (level anchor) Cardi B (radical:woman) at your favorite bar (memory palace), wearing a weave(radical:thread), sitting on a stool (radical:stool), drunk (radical:drunkard), talking crap (radical:sound) (radical:broken heart) about how men are dogs ど。She’s eating oreos (to weave:おる) and just rapping loudly(lookalike:say), but how discerning(しき) could Cardi B possibly be of people anyway? She’s so angry.

識:しき discerning
織:しき weave
織る:おる to weave

I personally write out my walks through my memory palaces. I color code my radicals, kanji, vocab, hints and lookalikes to not to distract from the story. When I do my Reviews, if I can’t remember something, I return to the memory palace first in my head and then on paper if it’s not a strong enough of a connection. Then I make some tweaks accordingly.

It’s a lot of work. I’m not going to lie, but it is well worth it. Eventually you will only spend a few minutes writing up your memory palace step by step every four to five kanji. In the long run this is much better than being forced to see that very same kanji…over and over again before you get it….twenty or so vocab waiting in the wings until you have finally mastered it.

About the book:

About my favorite mentalist Yan-Jaa:


There have been a number of discussions about these methods in the past with different methods and approaches, mostly to deal with the shortcomings of WK’s default mnemonics. For example, here’s a thread on my approach for readings (I don’t have trouble with radicals). There are links there to several others.

It sounds like your approach is a lot more complex than what I’ve done, tying into the whole level rather than individual kanji and vocabs. Maybe that’s needed when you get up to the higher levels. I haven’t hit that point yet, so I’ll be sure to check this out some more when I get there if I need to adjust my methods.

1 Like

I came to this topic thinking:
“Great yet another ‘I know better’ topic that probably won’t have much thought behind it :roll_eyes:”.

Instead, I learnt something today. Thanks for the fleshed out advice to help us augment our experience.

I do have one question though, which I will preface with some context:
When I learn the Kanji, I do my best to drop the mnemonics by the time the item levels up to Guru. This is because I hate having a “checkpoint” between the relevant data points (Kanji visual representation, meaning, reading).

So here’s the question: How easy has it been for you to drop your memory palace and directly access the data you require as quickly (or close) as you access the word “cat” in English?


I haven’t. Which is why I asked the question. Unlike you, I don’t believe I have everything figured out. I figured you’d be interested in answering the question sans arrogance. I stand corrected.

I think I will skip getting back to you altogether. I have no real interest in feeding your ego.

Congratulations :tada:

How right you are. I should be more humble, like you. [/sarcasm]
Get over yourself.


Are we on a different forum or something? I can’t see where any of this happened. He asked an innocent question and you answer with a condescending rant?


I like where this is going, but I forgot to buy popcorn while doing groceries so…



6 months? I’ve been working on memory stuff for years.

In any case, I don’t think that’s going to be my problem, when I get there. I have a very visual memory. Also, I did Remembering the Kanji years ago. Despite my memory being a sieve, meaning I’ve forgotten a lot of the kanji between then and now, relearning the shape of the kanji’s form is generally pretty easy for me. Meaning is only a minor issue. WK uses different radical and kanji keywords, but since I’ve forgotten most of the RTK ones, it’s not a big issue most of the time, and synonyms are great where it doesn’t.

No, for me, it’s all about readings. And pretty much has been from the beginning. WK’s mnemonics are pretty terrible in my opinion, because they’re more than willing to use the same mnemonic for different pronunciations. It’s insane. Why would you use the same mnemonic for 3-4 different things? “Cow” should not translate to any of む, むい, かお or うし depending on WK’s whim. Madness.

But I do a reassess every 5 levels or so and analyze my performance and what problems I’ve been struggling with. Any pain points I try to address by changing my methods. If I get there and I need to, be sure I’ll try your method (as well as 6 or so other things which probably won’t work). :wink:


Okay, I read your posts like three times, but this really just stood out to me. I don’t know about others, but I’ve never had this problem before.

Here’s why I find it odd. Every level, I just power through my kanji reviews — I usually use the mnemonics, but I always pass them to Guru even if I get them wrong so I can reach the vocab.

Yes, I’m cheating. But when I do get to the relevant vocabulary, it reinforces the kanji readings and meanings for me. After leveling up, I no longer have to worry about getting x number of Kanji to Guru and can let the SRS system do its trick.

And it works. I’m not being cynical or judgmental here — I just want to understand why making these elaborate constructs is necessary in the first place. When you eventually get to reading native content fluently, you will invariably have to discard mnemonic devices, because you will have to immediately recall the meaning and reading from the shape of the word, not from these devices you construct.

I’m aware everyone has their own approach to studying Kanji. So I’m not saying this is a bad method — it may well work for many people. But if there’s another method, which consumes less time and is easier to use, won’t some of you guys consider it? :wink:

Which brings me back to the original point: not being able to recognize vocabulary until you recognize the Kanji. I don’t know about others, but for me having more contextual vocab helps me remember the meaning and reading. So the effect is kind of the opposite for me, I guess. Above Guru I’m pretty sure most people stop using mnemonics anyway, so I find it more efficient to make them quickly.

Disclaimer: I don’t really know anything about neuroscience, skydiving, anthropology, genealogy, or environmentalism. I can play some chess, though.


And then you dropped in out of nowhere with a rant about his rant. :expressionless:

I deal with this in a similar way - vocab and context is where it’s at. For a couple of previous levels I had to battle my way through getting kanji to Guru, because neither the story mnemonics, nor the phonetic mnemonics would work for me and I started seeing kanji I’m not yet familiar and/or comfortable with. This later gets sorted out once I get to learn the vocab items.

When I struggle, my tactic is usually the following:

  • repeat the kanji cards around 3-5 times in a random order with a short break in the middle (say 2-3 minutes?) shortly after doing the regular WaniKani lesson pile. This usually gives me fairly good retention and later the WaniKani SRS does its job.

  • for kanji which don’t stick I add more (useful) words to my Anki sets with the kanji in the first, middle or last position

  • add synonyms to the WaniKani cards to work around SRS limitations (f.e. oh it was not “rice field”, it was “rice paddy”)

Unfortunately, since my visual memory is weirdly selective and works only for specific details, I’ve always been developing and relying on echoic memory which for a language like Japanese is deceptive, but works if one fine-tunes minor differences (hence why WK’s phonetic mnemonics don’t work for me) and relies on larger context blocks. The upside is that I much easier associate emotions with words and internalize them :slight_smile: .


Same with me, most wanikani mnemonics that I remember the best are the ones that make no sense to me, because then I associate the emotion of frustration/confusion with it and remember it better lol


Not sure if we have a different definition of a rant, because that definitely wasn’t one. It’s quite lacking in length. Nevertheless, I think it was justified.


Alright, it was probably justified, it was just a little uncalled for, given that the two people in the argument were already debating it completely unrelated to us. Also, “quisitive question” is extremely redundant. Using bigger words doesn’t make your points more valid. Not to mention that quisitive isn’t really a word.

1 Like

Maybe, but given the user’s behaviour in multiple threads I have a hard time regretting it. I can agree it was a bit petty.

Well this at least was uncalled for. I was trying to emphasize that @VegasVed was asking an innocent question without any premise; not “use big words”. Maybe I chose the wrong word. Is inquisitive or innocent better?

edit: cleaned my initial response to OP a bit


you’re desire to use big words is quite antidisestablishmentarianism and I’m personally quite offended by the prevalence of pneumonoultramicroscopicsilicovolcanoconiosis in your posts.


Well, I can agree that the WK mnemonics are often hit or miss. Maybe that’s why I recommend the community mnemonics script so often :slight_smile:

But back to the recognition point — for the more common or distinctive kanji you can use minimal pairs (I don’t think there’s terminology here) of vocabulary to differentiate them, such as X kanji in XY word. However, at a certain point you start running into Kanji that are really only used in a few common words — I believe a good low level example was 努?

One can imagine by thinking of them in the same vein as hapax legomena that it would be more difficult to remember such Kanji. However, for some weird reason this doesn’t really happen. I distinctly remember this effect a few levels ago when I had to remember the kunyomi ほたる, はたけ, and はち (蜂), but because I was in a hurry I didn’t use the mnemonics.

If we’re advertising our mnemonic methods :high_touch: sometimes I just remember the pattern that I have to type the kana in, sort of like how some android phones can unlock. Add that to your phonetic memory and you get even more accurate results.

Also @morteASD, perhaps you’re looking for “a question asked in good faith”.


Okay, this isn’t even opinion anymore. I know there’s been a lot of metaphors for brain functions and mechanisms of memory being thrown around here, but I highly doubt that this is the case.

Why? Because it’s just an unprovable and unfalsifiable statement; really, how would you be able to tell that this is the psychological mechanism underlying memory?

Seriously, a lot of this is strongly disputable. “…eventually [they] fade into your subconscious”; what does this even mean?? Is it even useful to make the distinction between not needing mnemonics and them fading into the subconscious? How relevant is this information, even? It’s not like it impacts the actual memorization itself.

And that’s if it can be accepted as true in some sense, which is questionable because there is no applicable model to base this whole thing on. How do you know learning a language uses the same underlying mechanisms as memorizing factoids, learning physical skills, or even playing an instrument? And with what methodology can you answer these questions?

If you cannot reasonably test your hypotheses, if you cannot logically falsify them; at that point, even if the original proposal is true in some sense it’s not even basically scientific anymore. You cannot use terminology without first defining what they mean, and you cannot apply your own framework of understanding or your own model to your claim without first presenting said framework or model!

Logically, I don’t care about the way you present your arguments as long as they’re scientific and pertinent to your core ideas. I had a similar annoyance with refold, which for the life of me I couldn’t find any backing science for, and which has similar claims about long-term efficacy. However, at least it recognizes its own limitations. Your posts seem to indicate that you cannot even do that for your own methods.


Wow. What a nice condescending rant!. “My husband is chess master Chad Morrison”?
“My rate of accuracy is 98%”…just wow. An extremely patronizing and brag filled post. If your way works for you great. But your post just tells everyone that you are a god of kanji and your way is the true and only way to be successful at this.


You forgot the part where she decided to throw in the fact that she’s also a female paratrooper, because of course, that’s totally relevant to a conversation about kanji knowledge acquisition.

“I jump out of planes and my husband is a chess god; checkmate, kanji enthusiasts!”


…how’s the JLP N1 prep going? God, it’s as if you jumped into the deep end without even bothering to check how deep it was!

Idk, is this petty? But to me it almost seems like a serial troll more than someone with realistic goals.