100,000 Kanji writing Reps: My thoughts

Revision Kanji i’ve written as of 10:00 a.m today for the last few days of newly memorized Kanji

Today I’m going to hit 100,000 written repetitions of Japanese Kanji as I head into the homestretch of about 200 Kanji left to wrap up the 2,136 Kanji for the Jouyou list that the government says is the base for most required reading. Like a lot of people i’ve been having many challenges with life, but I wanted to make sure I made an update toady as it is a pretty big milestone.

Where I am Now

This is a long post, so I’m dropping this here first. I’m hovering around 1950+ memorized. What this means is that for all 1950 of these Kanji I can read the Kanji’s Onymi reading (some Kunyomi) and I can write the Kanji from memory based on the Keyword. I can read quite complex chains of Kanji/sentences with few issues. This does not mean I know all the words in a sentence (this will be mastered in the vocabulary phase, which is phase 2) My reading speed presently relative to whatever I’ve been exposed to. Meaning, If I encounter a new word in an example sentence I can read it slowly, but correctly. I have not been reading anything at length  (that’s phase 2) and therefore have no expectations for reading with speed (yet). But I am reading much faster than I did in the beginning. The dots have started to connect themselves. The door has been “opened” as it were, to really put pressure on the other elements. By being able to read almost anything I see I can start to dial in on all the other elements. Okay, on to the post.

The Journey of 100,000 reps

In my last post I spoke about something called “Basic Equivalency”, which is the very binary viewpoint that what you put in is what you get, particularly if you ensure that what you put in is highly efficient and guaranteed to give you a better return on your investment of time and energy.

People will “study” for several hours a day but cannot accurately quantify what this study is, where it is leading them or how it actually impacts their overall progress relative to a certain goal. In writing novels, learning languages and tracking other tasks I’ve always found it unusually important to be as drastic as possible when tracking one’s progress, because only through tracking do you know where you actually are. I’ll be going into why I focused on writing and my results as I continue and to let you understand that I am not doing rote memorization, but what I call “reinforcement”.

You see, the knowledge of where you are is fundamentally important when you have a very large strategy you are implementing with many moving parts. I found that what can become overwhelming with the study of anything complex, is mostly the perception that it is complex, not the actual complexity of the problem. This means that the less complex a large task appears, the more likely you are to  accomplish it. 

Therefore, the more complex you perceive your task to be, the more stress and unwillingness you will have to finish it. 

Note: This does not mean your task is NOT complex, but you must ensure to make it less complex (perceptually) to the point where you are just focused on implementation and understand exactly where your efforts will take you.

In the past when I tried to learn the Kanji, I did a lot of Anki repetitions, ran through James Heisig’s Remembering The Kanji book and generally felt as if I was on the right track, but like most learners I was not. We think that with an application designed to help us memorize things we will “memorize them” and therefore have an easy route to language acquisition. But the nature of the way the brain works and how we actually store data has lead me to believe this perspective of acquisition is a complete illusion, which I will elaborate on later. Remember this: Asian languages are perceived as ‘harder’ to learn because they have a massive barrier of entry. Unlike German, French, Italian and Spanish (where the average native English speaker knows thousands of words by default and need only learn pronunciations to start quickly reading and using the language) Japanese has no similarities to your native language. There is no way around needing to be able to read the language to truly progress in it. You cannot learn words without being able to read them, nor can you really study grammar heavily without being able to read the grammar. People would say otherwise and point to ‘kids’ (I will address that). but for self-directed study that you will be doing in a short period of time I think its better to get this major barrier out of the way so you can progress far more quickly towards focusing on mastery of the language (without focusing on the massive reading limitation). Almost all problems one will have in the future will be relative to one’s ability to read and process the language. Simple question: Is there a downside to being able to read all the Kanji you see? Is it more of less likely that this ability will drastically help you make massive gains in your attempts?

I like posing these types of questions to myself as I did with German. I said to myself, if I can read and recall 4,000+ German words it is more or less likely that my German will improve? I think you know the answer to that one.

Here’s why I focused on writing:

Why 100,000 reps and writing?

The internet is filled with people who talk about methods to learn read and write Japanese. Many even are quite adamant that “writing is useless”, “you learn Kanji from words not Kanji” and all other kinds of stuff that frankly i’ve found makes no sense relative to the true acquisition of reading and writing ability. This argument has never held any weight for me, simply because all Japanese people are required to learn to write Japanese Kanji. 

Since this is an indisputable truism, it means that there must be value in the written process as a part of acquisition. Ignoring this relative to Kanji, I think is a fatal flaw in utilizing certain strategic approaches.

The value of writing is relative to a few factors: mainly it is a kinetic process and your muscle memory directly links to neural networks that you build with your knowledge of Kanji and words. However I am not speaking about Rote memorization, (which is no issue if you are learning at the pace of the average Japanese child).

A quick look at how kids in Japan learn

Japanese children have a very leisurely ride towards Kanji acquisition. Students generally take 10-12 YEARS  to learn the 2000+ Kanji. Take a breath here.

Rough math points this to around 1 Kanji per week, for twelve years. much of this learning is done through rote repetition (writing) and then natural exposure. So it is a nice, leisurely twelve year ride to developing strong reading and writing ability.

As adult learners we do not have the luxury of 12 years, therefore our methods must be highly efficient and effective relative to our use of time. Here’s why writing has been extremely efficient, based on what I know about brain science.

Rote memorization works if you are focusing on one Kanji a time per week for 12 years, because you’d be essentially mastering handfuls of Kanji at a very slow pace over a decade or so. This means that if you drastically shorten this timeline, your ability to differentiate what you learn descreases massively, and your stresss levels go through the roof.

Why This Isn’t Rote.

I learned 1009 Kanji in one month, which means based on our school kid example, I compressed six years of learning time into one month. The method i’m using allows me to memorize the Onyomi readings before learning the Kunyomi so I cannot claim 100% reading ability relative to every iteration of the various readings (but i have learned many of them already and it has not been difficult one the first reading it mastered, but more on that later).

The brain is a fascinating mechanism, once used properly. In learning German, I dove very deep into the acquisition of vocabulary first which changed EVERYTHING, because vocabulary is basically the cellular structure of all language. You can’t do anything without words. Similarly, when it comes to Kanji I came to a similar observation. Romance languages have easy access for a new learner because we are able to read them immediately. The “building blocks” of these languages are the romanic alphabet systems which we have already mastered. When we look at Japanese, it is natural for one to assume the “building blocks” of the language are the Kanji/Hiranga/Katakana, but they would be wrong.

The building blocks are literally first, the primitives which build the Kanji, THEN the Kanji themselves. 

This makes your approach much easier, because if you know all the building blocks, then you can technically write and recognize all the Kanji. This is the basis the powerful work like Remembering the Kanji by James Heisig, where I saw this implementation for the first time.

There are about 200 primitives that build ALL Kanji. So when you think about it, being able to write all 200 primitives means you can write ALL the Kanji that ever existed (even if you do not know them). Additionally, by solidly learning the primitives, you are guaranteed to have no issues writing the Kanji, and more importantly, differentiating similar Kanji over time. 

Heisig’s work, though ground breaking, still felt “jumbled” to me, because it is logical in its approach, but it wasn’t as organized as I’d like it to be. I could learn how to write Kanji but I had no idea how to solidly learn the various readings (or eventually, how to avoid mixing up similar Kanji). This is a problem that happens with very large data sets and Heisig did a fantastic job opening the door to methodologies of acquisition, but it was hard for me to see the “end point” after I was done. What was the approach?

What I’ve Done.

I approach what i’m doing now like physical training that requires different kinds of approaches for different situations. First I will speak about what I call reinforcement. 

like Heisig I cam upon some interesting research by a person named Alyks who proposed using movies as loci for groups of Kanji. At the time I had just struggled through Remembering the Kanji, and was thoroughly stressed, confused and unwilling to restart the entire process.

In recent years, as I’ve done quite a bit of heavy research on memorization, methodology and acquisition, my results with German made me look back on Japanese with new eyes and I understood the value of  movies and characters as loci for entire groups of Kanji. Whereas people normally try to turn every individual Kanji into an independent story, my target group is linked to one primary loci (movie, character) and therefore sits in its own container.

The Problem: If you are attempting to learn quickly, it is extremely hard to learn Kanji via individual stories with your own creative imagination, as there an unusually high potential for extreme confusion, overlapping images in various stories and being unable to properly differentiate nuanced meanings for words like “receive” and “accept” and the high likelihood of eventual burnout, as it requires incredible amounts of creative energy.

Solution: Kanji that lie in their own containers are in their own “universe” replete with its own images and continuity (and previous memories you have) and drastically reduce any potential for confusion. Also the results of your efforts are far more substantial relative to the same mental demands. Each story I create for each Kanji is based around these “movies, series or characters” within the context of that loci. For example, I memorized shuu (しゅう)with the movie ‘Terminator’. Terminator 2 has a lot of a SHOOting that happens. 

So for the 25 or so Kanji for SHUU each story is linked to a scene I create from the movie itself, or something relative to the movie. I’ve watched Terminator 2 more times that I can count, which means the creation of the story (with the primitives) becomes a mostly visual exercise.

Stay with me here, i know this is a long post. When I write these Kanji, with each written repetition called a “rep”, I reinforce this story. I revisit the visuals, I revisit wherever I was at the time, and I revisit kinetically the feeling of the story that I created. This creates what I called “multimodal” memorization, because I am implementing multiple senses as I learn these stories, which virtually guarantees recall. If I was simply looking at Kanji on a screen and trying to figure out which was which, I would have only one context (visual) to work with, which means i’d have to see a Kanji hundreds (perhaps thousands of times) before my brain properly memorized it. Multiply this one Kanji by several thousand and you are presented a massive problem, which I call the “endless revision whirlpool”, where even though you’ve supposedly ‘memorized’ all these Kanji, you must constantly revise them. True memorization equates to 100% recall, without the need to revise. 

Its the reason you can read these sentences without the need to look anything up. You’ve memorized the alphabet and have no issues reading any of what i’m writing. That is the goal.

Each time I learn a Kanji I have around six or seven stimuli associated with the memorization, which link to a pre-existing long-term memory (movie/character/series). So i’m tapping into well-established memory networks deep in long-term memory and adding different contexts. This ensures a certain kind of recall which quickly becomes a long-term memory through massive reinforcement.

Each time I write the Kanji, I make the story stronger and the reading more permanent. I am not writing to recognize the shape of the Kanji, I am writing to reinforce the logical story already in my mind. So you develop: spatial recognition, kinetic impressions, visual contexts from past memories plus the new contexts you’ve created.

Why this is extremely important: many of these systems present very smart ways to approach learning large volumes of data, but tend to give no guidance in how to truly memorize them for practical use in real life. 

Remember: True memorization equates to 100% recall, without the need to revise. 

The end goal of all language learning is recall. No mater how good the system, if I cannot recall what I have studied quickly, then the system has failed me.

Why 100,000 part two

Okay so now that we understand that, it is important to understand what a ‘rep’ is.

Each time you write a Kanji, you are retrieving the story from your mind. You are revisiting whatever you came up with and putting it deeper into long term memory. In fact, after a while with many of the writings, you will forget the story and just know the reading because the brain simply recognizes the Kanji as is. There is no shortcut for practicing these retrievals. This is what rote memorization is, a forced practice of retrieval until you store the word in long term memory. But constantly retrieving a story from your mind (with multiple stimuli attached to it) is like adding gas to a fire, you are simply making the Kanji stronger and stronger with each writing. Also: After the initial learning session (I write the new Kanji 20-35 times) you only need to write the Kanji once per day as part of revising the group. This memorization happens pretty quickly. I wake up early the next day and rewrite my newly learned Kanji and generally I have 100% recall of all the Kanji. Sometimes one or two I might not remember a primitive, or I made a part of the story too vague. But generally, its 100% which is insane.

In the past, this is where I would stop. I would “believe” i’d memorized the Kanji and move on. However, I know it is only through repeated exposure that you ensure long term memorization. The process i’m developing will only take 8-10 weeks. You want to ensure you don’t need to really do much revision beyond your acquisition phase. 

There is learning. Then there is mastery. I’ve found that aiming for mastery sets up a different stage, mentally and psychologically. Initially I set a minimum of 800 reps per day, but eventually landed on 1,000 reps per day since the number is round and didn’t take that much more effort. These reps are a mixture of reps from newly learned Kanji (20-35 individual reps to reinforce the new story) then several hundred revisions of previously learned Kanji. This process takes around 1.5 to 2.5 hours per day I think. To some this might seem extreme, but we must ask ourselves a question. How do we GUARANTEE the likelihood of a certain outcome?

A high level of comfort reading these Kanji (I believe) cannot come from writing 20 per day and skimming through Anki cards. Japanese kids master these Kanji overtime. Mastery only comes from high focused effort maintained consistently. Mastery is also incremental, meaning basic steps are mastered before progressively harder ones. I knew that doing thousands of repetitions guarantees results, and 1,000 per day seemed to be a good number for mastery practice. It is high enough to seem intimidating (initially), but not low enough to feel lazy. I can also track my gains directly by the numbers. Remember, the major barrier of Japanese is reading and there is no way around it. The ‘intensity’ of approach is relative to the outcome one desires.

Final Thoughts

It is VERY IMPORTANT to remember here that what I’m doing is compressing a 12 year learning period into 8-12 weeks. This is lightning fast by any stretch of the imagination. To give you a sort of “top down” view this is how it looks using 10 weeks as standard.


10 weeks – 2,136 Kanji, 297 readings, 1,000 words (free time).

so for PHASE 2 my focus will be heavily vocabulary based. I will save myself hundreds of hours in revision time because if I can read everything I see, my focus is only on the meaning of the word. This is why a short window is important. You technically do yourself a “disservice” by moving slower. You will constantly hit obstacles that are reading based throughout your journey(which have the potential to discourage you to the point of stopping dead in your tracks). In fact, I daresay that almost every obstacle you face in the start of the journey is vocabulary based. Even with a romance language, if you can’t read what’s in front of you, you can’t use it. In fact, the same things happened for me with German. After I crossed the 3,000 word mark, I could usually read 90% of what was in a sentence, so I was always reading a sentence for the meaning, not individual words. It drastically changes how you are able to absorb the language. You start looking at sentences as entire partners that you master, versus individual bits and pieces you are trying to put together. As you increase this exposure, the patterns start to make sense (based on how the brain works) and your comprehension starts to massively increase on its own. So again, there is no downside to a massive familiarity with the Kanji. 

I want to massively internalize these to such an extent that when I start reading and go into phase 2  I am not stuck in the “revision hole” where I keep having to revise all these Kanji I learned while trying to learn new things. It is better for the words and readings to naturally reinforce themselves as you now aim for thousands of contextual exposures. I have a small anki deck i’ve used to slowly revise new words and I’ve been pleasantly surprised at how easy the words are to read now. It changes the entire game in terms of training memorization. That extra energy to try and remember the shape of the Kanji (that you don’t know) is just a pain.

Where I’ve Been

I think its important to note where i’ve been. I have never had this level of success with Kanji.

None of what I’m doing I consider “magic”. It is extremely strategic and highly focused execution of a strategy that I knew would required tons of energy, godlike patience and a rock solid tracking system. I based a lot of what I was doing on a few years of incremental research and particularly my observations with my breakthrough in German in 2020.  I had to project months ahead and assure myself what I was doing would work out, because I’ve been in situations where I’ve spent hundreds of wasted hours on inefficient methodologies which sucks terribly. In these journeys you must always:

Know where you are going: Timeline, specific goal within that timeline (with leeway for delays and life)

Know what you ‘should’ get: Know the specific gains you are targeting and based the majority of your observations on that metric.

Know your next step: Know how this connects to the next logical step and how it makes that step much more efficient.  – so using this as an example, I know that being able to read all 2,136 Kanji will give me the ability to properly start the process of being able to eventually read ALL media based text, images and the majority of Japanese written material (with the exception of novels). This means that my vocabulary acquisition phase will be much smoother as I will have (a) reading confidence and (b) much more effeciency in recall (c) more time to practice memorizing readings within context versus trying to remember Kanji + reading + meaning. This ability will then lead to — enhanced ability to practice speaking and go towards relative fluency based on the number of hours of listening and vocabulary acquired.

Okay, i’ll see how things proceed. At this point I will start to design the learning program based on all these methods .


About marcusbird

Writer, Designer, Filmmaker
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s