Review 1: Remembering the Kanji, vol. 1
2 months, 520+ kanji later, here are my thoughts on learning kanji and primitives through Heisig's classic textbook series and a related Anki deck.
Kindle says I purchased Remembering the Kanji, vol. 1, on July 6, 2022. I don't exactly know when I started using it in earnest, but it must've been shortly afterward. Today is September 3, so I'm reflecting on my experience of it plus a related Anki deck within the past two months.
I started relearning Japanese around mid-March of this year. My main study tool was Fluent Forever's program, which included a mobile app that has a Spaced Repetition System and vocabulary frequency list, as well as weekly coaching sessions to practice listening and speaking with a native Japanese speaker and trained coach. However, I realized over time that as I added more flash cards to my FF SRS, the kanji within them were getting harder to remember and distinguish. I truly only was memorizing general shapes of kanji and how they appeared in specific sentences in the app. I was not remembering specific makeups of kanji, which I thought was OK at first. However, over time I found too many collisions of similar-looking kanji, and I realized I was just tricking myself with getting familiar with a flash card rather than the kanji itself.
I decided that I needed a different tool dedicated to learning kanji to supplement my Japanese relearning. I eventually settled on RTK.
After less than two months with RTK, here's my progress:
- 39% or 190 out of 490 pages through. Note: pages 437 to 490 are indices.
- 535 out of 2,200 kanji encountered, not including primitives
- Midway through lesson 19 out of 56
Paired with this Anki deck:
- 242 kanji cards are mature
- 285 kanji cards are young+learn
- 1673 kanji cards are unseen
- 0 kanji cards are suspended or buried
I have set my pace in Anki to learn about 10 new kanji a day. Initially it was a little higher (I think the default was maybe 15). Factoring in fuzzy discrepancies like that, I must have been using RTK for the past 53 days, give or take a day. This pace has overall been healthy for me, as it's enough to make sustainable progress without being overwhelming. At this pace, I would finish RTK vol. 1 in 220 days (7.33 months) total, currently with 167 days (5.56 months) left to go. Other folks somehow claim to have done 20 to 30 new kanji, so they would in theory finish in half or a third of the time. That sounds so incredible! But I don't honestly have that time or stamina.
Since graduating from college 10+ years ago, several kanji-learning techniques have become available. I'll highlight a few popular ones I've tried.
There are many books. I already owned Genki as it was my college textbook series, and since I felt that didn't give me much at all in terms of kanji study, I did not bother revisiting it. Genki combines grammar, vocabulary, listening (with CDs), and kanji, so it doesn't specialize in dedicated kanji study. That makes this series good for formal courses where teachers typically want to approach Japanese for students by helping them learn writing, reading, speaking, and listening all at roughly the same time from the beginning. Now that I'm guiding my own learning, from all the things I've read and watched on Japanese self-study, I became convinced that I want to use kanji-learning-specific tools for kanji.
Wanikani is a web app that came out in 2012. My coworker who is also learning Japanese has used this up to lesson 10 (?) and has really liked it. From what I remember, Wanikani orders the kanji by JPLT proficiency level from N5 to N1 (beginner to expert), provides mnemonic stories for kanji meaning (using a particular English keyword) and (onyomi, kunyomi) reading, and has a built-in SRS system where you type in answers. Apparently there's a vocabulary piece as well. To keep you consistent, the system emails you reminders, and if you miss days, the review work adds up, which can be pretty demoralizing.
The first 10 lessons (containing a few hundred kanji) are free and will occupy you for awhile, so it's not a bad deal. After that, it's $9 a month, with other pricing for annual and lifetime.
I very briefly tried it out and quickly got frustrated. I thought the mnemonic stories were too overwhelming for me – while combining meaning and reading sound good (no pun intended), that combination maybe contributed to me feeling overwhelmed. Finally, the interface was a little difficult for me; it was easy for me to make typos, which would prolong my review time in an arbitrary way. I probably didn't try this method out long enough before dismissing it, though.
I briefly tried the Kanji Study mobile app as well, which is mostly free but has some paid features. I coincidentally learned that TokuyuuTV used this app as a primary tool to learn kanji. He is now fluent, living in Japan and working for a Japanese company. Kanji Study's a very polished app that lets you learn kanji in several orders of your choosing – there is even an order for RTK. I tried the default one that is in order for JLPT proficiency.
The Kanji Study teaching method seems to be to teach you multiple meanings and readings at once, using flash cards with sentences that use the kanji in different ways to help reinforce the multi-faceted meanings and readings for each kanji. This method did not stick with me – I found it overwhelming and not effective. TokuyuuTV said he needed to create his own mnemonic stories to use this app, and maybe I'm just too lazy and not creative enough to do that. :)
However, I still keep this app on my phone. I occasionally use it to get deeper information on specific kanji, such as animated stroke order for review, etymology, and the evolution of a kanji over hundreds (or thousands?) of years from ancient China to modern Japan. It conveniently links with my mobile dictionary app of choice, Takoboto.
The last kanji-learning technique I'll mention, of course, is the textbook series Remembering the Kanji by James Heisig, which originally (actually) came out in 1977, but the latest editions came out (also) in 2012. Like Wanikani, it has mnemonic stories and English keyword(s) for each kanji and primitive, but there are a few crucial differences.
First, the order is not based on JLPT proficiency levels or frequency lists; it's based on "primitive" grouping. "Primitives" are Heisig's term to describe pieces of kanji that have their own distinct meaning and build upon each other to create many kanji. I recall fuzzily that they are technically different from the official list of radicals, but to me they are pretty similar in concept. A downside to this grouping order is that you end up learning some very rare kanji early on and some very common kanji later on. An upside is that I've found it incredibly effective.
The second crucial difference is that RTK volume 1 (and maybe volume 3) focuses only on meaning, while volume 2 focuses on reading. These differences have made RTK very controversial in the Japanese language learning community.
Why did I settle on RTK? It's a bit arbitrary. A month or so ago I became obsessed with Matt vs. Japan's Youtube series. In particular, I really liked his video from 2018 called Why "Remembering the Kanji" is the Best Way to Learn Kanji. (Ironically, in early 2022 he claimed that he completely disagreed with the video in favor of a competitor product he was selling, but I took a look at the product and wasn't convinced...) What put me over the edge was his interview with this person Stevi who became fluent in Japanese, and he mentioned his method of learning kanji in 2-3 months was via RTK. Like the primate that I am, I became convinced that if it it worked for this guy who is now fluent, it should work for me!
Finally, while I frequently read in Youtube comments and Amazon reviews a lot of hate for RTK – how stupid or tedious or ineffective it is so most of those naysayers quit midway – occasionally I would come across posts by people who actually finished RTK. All of them (with one exception) said that the struggle was completely worth it.
Bonus: today, I happened to come across this NihongoShark blog post advocating for RTK, so I feel further validated in my choice.
My learning process
Here is how I learn about 10 new kanji a day, not including primitives in that count. I have a morning and evening routine:
- In the morning, I carefully read through and digest each primitive or kanji entry (with keyword and mnemonic story) in RTK.
- I draw out the kanji or primitive in my physical notebook with a pencil.
- I write the keyword in blue ink.
- I write my version of the mnemonic story, typically a shorter version, in black ink.
- I underline the keywords from the kanji's primitives that appear in the story.
- I star primitive keyword meanings when applicable.
- If I have time, I review the new kanji immediately by drawing them from memory in my e-notebook.
This takes about 10-15 minutes if I am focused; another 5-10 if I am not. It is fun since I have an excuse to use different colored writing utensils. :) Time passes throughout the day. I theorize with my pseudo-neuroscience knowledge that this short term memory starts to stew around in my head. Plus, I don't have time to do a ton much else in the morning before work.
- In the evening, I open Anki and my RTK deck on my phone. I have my e-notebook next to my phone.
- I test my knowledge on a kanji by pulling up the flash card question (the English keyword) in Anki, drawing out the kanji on my e-notebook, then comparing with the flash card answer (the kanji). If it's super close to perfect in terms of what I intended to accomplish, I say I pass it. If I missed fundamental parts or truly confused it with another kanji, I say I fail it.
This takes about 30 minutes now, given how large my deck of mature/learned/new cards has been growing.
It may seem redundant for me to have a copy of the kanji, primitives, and mnemonic stories in a physical notebook, along with the Kindle textbook and the Anki cards. However, I've studied this way in school and taught myself programming for my career this way for many years. I may be stubborn and inefficient, but I know this way has been effective in helping me solidify concepts in my head. Plus, maybe this is a price I pay for getting a Kindle textbook, and I secretly want some physical version...
On RTK Effectiveness and Positives
I have found RTK incredibly effective so far in quickly learning kanji meanings and keeping them straight. I very much echoed Matt vs. Japan's sentiment in Why "Remembering the Kanji" is the Best Way to Learn Kanji about learning the specific parts of a kanji as opposed to fuzzy outlines of them that I would only get from FF flash cards. When I encounter kanji I've learned via RTK in real Japanese-native content, I can accurately guess their meaning. When I encounter them in Anki or FF flash cards for vocabulary with hiragana, I can more easily remember their meaning and distinguish them from other kanji in other flash cards. Reading with onyomi and kunyomi is still a little hard, though. Nonetheless, picking up meaning accurately in content and flash cards is so satisfying. I can't stress enough how satisfying it is.
Beyond effectiveness, I've been super happy with my cost investment. I bought the Kindle version of volume 1 for a one-time fee of $9.99, and there are many free Anki decks modeled off of RTK (this is the one I picked). Of course, Anki is great itself because it's free, not tied to a subscription/proprietary platform, and can be used offline (but, I have donated to the OSS folks as a sign of appreciation!). I plan to get RTK volume 3 (skipping volume 2, per recommendation by Matt vs Japan) at a similarly cheap price.
It's easily motivating for me to keep up with RTK and Anki for these reasons:
- The fear of missing out on Anki review each day is real. For the first time, I missed 2.5 days of review for RTK Anki, and it took me a good 3 hours or so to catch up. It sucks.
- Input/feedback is super easy and not prone to typos, as detailed in My learning process above, compared to Wanikani.
- I don't have to make up mnemonic stories myself from scratch! This is a huge barrier lifted from me. I can still tweak them to my own desire if I wish. The particular Anki RTK deck I downloaded has its own, often less problematic (see On RTK Difficulties below) and shorter mnemonic story alternative too that helps.
- By focusing only on meaning, I'm not overwhelmed with needing to learn reading at the same time to pass a card.
- Writing is fun! This way of reviewing is both low barrier and enjoyable.
- Writing is effective! I've heard it's a waste of time if you just mainly want to understand (reading, listening) and speak, but again, given how kanji can just be fuzzy pictures if you don't truly take the time to learn specific primitive makeups, writing helps so much here. Bonus: writing is a nice skill to have if you ever need to fill out forms, learn 書道, or just want to show off to friends.
On RTK Difficulties
Positives out of the way... my feelings seem to be mirroring this Scope of Knowledge graphic from my code therapy post. In the first 200-300 kanji (month 1), I felt pretty awesome. Already I probably had learned the meanings and particular drawings of more kanji in 20-30 days than I had in my entire 4 years of studying Japanese in college. Granted, I had a leg up in quickly memorizing certain kanji due to previous exposure in college (and elsewhere), so (for this and others reasons) by no means am I saying my college education was worthless. I do wish I took kanji more seriously and were self-motivated to do this kind of thing when I was still in school. Alas...
However, with kanji #300 and up until now, it's been getting pretty hard. I have more kanji to keep straight in my head while learning new ones. The hardest part for me is that many keywords I've learned are similar to new keywords I must learn, and some mnemonic stories have words that overlap with other keywords but aren't actually used as keywords in those particular stories.
For example, the kanji for fat 太 (#494), round 丸 (#44), and obese 肪 (#533) have different primitives and stories, but their keyword meanings are very similar.
Another example: silence 黙 (#255) has the primitives for black 黒 (#186) (which has the primitives for computer/ri 里 (#185) and fire 火 (#173)) and dog 犬 (#253). The story for silence 黙 says to think of a black dog named Darkness (for black) and then the song "The Sound of Silence" by Simon and Garfunkel that starts with, "Hello, darkness, my old friend." However, darkness 暗 (#519) is its own keyword for another kanji later on and has completely separate primitives!
Aside from keeping things straight, many of the mnemonic stories are just problematic. It's not super surprising coming from an old white guy (Heisig), on top of the fact that the kanji themselves must've been comprised together by old-fashioned, outdated, problematic (heteronormative, raunchy, sexist) norms from ancient times. I try to change up the mnemonic stories for me to be less problematic, but I am afraid of changing them too much such that future kanji stories collide with my changed stories from previous ones. Plus, problematic or not, a lot of the stories are so effective because heteronormativity and sexism are well-ingrained in my head, whether I like that or not.
Furthermore, some of the mnemonic stories and keywords are difficult to understand. Some of them reference older Western-educated material that many English readers today of different backgrounds and ages may not get. The example of Simon and Garfunkel is one such reference, and another is the kanji for decameron 旬, whose keyword and mnemonic story come from this same-named story from the 1300s I'd never heard of before... One the one hand, I guess I'm learning more English trivia, too! On the other, it makes memorization a little harder.
Finally, it should be called out that the time I currently spend on kanji alone may or may not be quite a lot for just 10 new ones a day (while maintaining 500+ in my head). I'm not sure if this method will scale over 5.6 months to eventually remember 2,200 to 3,000 kanji total. I may need to be more lenient with kanji I pass vs fail – if my goal is grasp kanji meaning generally well, rather than write perfectly – and maybe I need to skip the initial morning routine.
Overall, I'm glad I picked RTK over the methods I briefly played with. For the price, effectiveness, low barrier, and speed, these combinations have worked well for me. I'll report back later in my journey – maybe when I hit 1,000 kanji...
I'm not saying other methods are garbage and that RTK is the best, though. Truly for my own needs, RTK has been good for me, and I know many others, like my coworker, have found success with other methods.
I've been writing some briefer thoughts here on RTK and my Japanese learning journey in general: Japanese language (re)learning progress.