It’s been quite some time since I started learning Japanese, and as time goes on it gets harder and harder to remember the specific details. It definitely hasn’t been easy and I want to make a record of some of the things that went wrong along the way. This will be a “living document” of sorts, and I’ll be adding to it as I continue learning.
“Maybe I can learn Japanese after all”
Jan 13th, 2016
I’d been curious about the language for some time, but it had always seemed completely out of my reach. I’d heard, of course, that it was “the most difficult language for English speakers to learn” and all that, and for some reason I guess I must have taken it at face value and never dug any deeper. All that changed with Loremaster Nojah’s video series on the original Japanese translation of Demon’s Souls. His explanations of kanji fundamentals, loan words, and basic grammar, tied in with something that I was extremely interested in at the time (games by Japanese developer From Software), gave me the confidence and inspiration to actually give it a try. At the time, I could think of nothing cooler than to play a Fromsoft title early by importing the Japanese version ahead of the English release. With this goal in mind, I set off.
Mar 1st, 2016
I don’t remember how I discovered Tofugu, but the first trace I have of my Japanese learning journey is an email I sent to report a broken link on their guide to hiragana. I practiced with some web apps and by trying to read tweets from 根本凪, a singer I was into at the time. For all the mistakes I made along the way, learning hiragana (and then katakana) as my first foray into the language was definitely the right move.
After a couple of weeks of practice, I decided to move on. This was a happy accident; it would take months for me to reach a point where I could look at kana characters and instantly recall the sound, but if I had tried to drill them in isolation, it probably would have taken much longer (if it even worked at all).
The long kanji grind
Mar 22nd, 2016
Satisfied with Tofugu’s approach to learning kana, I moved on to their flagship product, Wanikani, as my one-stop-shop for kanji and vocabulary. This was where things started to go wrong, though I wouldn’t figure it out for a long time.
Wanikani is often criticized in “serious” Japanese-learning circles and I don’t think the criticisms are entirely without merit. I can say for sure that for me, the methodology was, at the very least, very sound. Breaking kanji into components to aid with visual parsing, and then teaching kanji readings through vocabulary, is the exact way I would advise people to go about the process even still. The price is a lot harder to justify, especially years later when there are many more great resources out there, but it wasn’t honestly an issue for me.
However, the biggest problem here was actually with me. I had never learned a foreign language before, and I had no real perspective on what the process would actually entail. Wanikani teaches over two thousand kanji, which sure sounds like a lot! My immediate reaction was that I needed to get through it as quickly as possible so I could “get kanji out of the way”. The course is advertised as taking “as little as one year” and I took this at face value.
This kicked off what I can only describe as an era where I would dutifully spend time every single day studying and reviewing kanji and vocab on wanikani, but never feeling like I was making any tangible progress. I took multiple “breaks” where I would avoid doing any new lessons and simply try to tread water, but it rarely felt like it did me any good.
Doing grammar wrong
July 3rd, 2016
I don’t remember exactly when I started using Textfugu, but I do know that this was the day where I paid for a “forever” license ($99) with the intention of starting my grammar studies in earnest. I have some notebooks in storage that I’ll have to pull out at some point, but I can say for sure that my grammar studies were few and far between. Even after two years, I still hadn’t covered all the (meager) material. I’ll have to write something about how terrible Textfugu is at some point, but for now I’ll just say that while Wanikani’s $300/life pricepoint is a bit steep, charging $99 for Textfugu is nothing short of robbery.
Whether the blame lies with Textfugu or with myself for not seeking out something better, the end result was the same: over the next few years, I diligently kept up with my kanji/vocab studies and let grammar fall entirely by the wayside. This was, of course, the exact opposite of what I should have done.
Misadventures with output
Sep 3rd, 2016
Despite having barely any practical grammar knowledge, I decided it would be a good idea to try out HelloTalk. I wish I still had access to my account because I’m sure I would be horrified at some of the stuff I came up with. Off the top of my head, I recall joining adjectives with と (naively assuming that it would be the same as the English “and” construction) and making the classic mistake of using で with 住む (the correct particle is に). I would refer to the dictionary when I didn’t know the words to use, and in many cases ended up using words that were uncommon or had a completely wrong meaning.
At this point I was approaching sentences as puzzles to be decoded. It’s obvious in hindsight of course, but at the time I had no idea what a waste of my time this would be.
For the next two years, I made zero apparent progress. I often wondered when I would start understanding the lyrics of the songs I was listening to on a daily basis, or the dialogue in the (English-subtitled) anime I watched regularly. It was two years of hard work where I was constantly convinced I was on the brink of a breakthrough, but it never came.
July 1st, 2018
This was the day I was invited to a discord server for people learning Japanese and English, and I cannot overstate how much of a game changer this would turn out to be. I didn’t really start interacting with the community until a bit later, but the skill gap between myself and the seasoned regulars was immediately apparent. It was obvious that I was missing something, I just needed to figure out what it was.
Doing vocab wrong
July 3rd, 2018
Throughout this month, I made a concerted effort to drill my vocab production on Kaniwani, which is a companion site you can connect your Wanikani account to in order to share data between the two of them.
I no longer believe that actively studying recall is beneficial at all, and I’m glad my Kaniwani stint was a brief one.
Trying (and failing) to read
May 5th, 2018
It was around this time that my copy of Persona 5: Dancing in Starlight and Persona 3: Dancing in Moonlight arrived from Japan. In the time that I’d been learning Japanese, I had fallen a bit out of love with the Dark Souls series, but fortunately the Persona series had come along to take its place as a Japanese franchise that I would like to try playing early. P5D and P3D weren’t exactly mainline entries in the series but I was excited to try them out, and I managed to do alright…for the gameplay segments, which barely required any Japanese knowledge at all. I was struck by one line in particular from one of the songs — さよなら言わせない — and how I didn’t even recognize 言わせない as a conjugation (this is how I learned that Textfugu didn’t include anything about causative form).
Aug 10th, 2018
If you ask for a good beginner manga, many people will recommend Flying Witch, Yotsubato, or Chi’s sweet home. Me, I decided I was going to read 虎子、あんまり壊しちゃだめだよ. I can’t even remember where I heard about this, all I know is that the premise (sweet misunderstood girl with superhuman strength accidentally destroys everything) was funny and it was relatively unknown and had no English translation. Needless to say, my initial attempt to read it was pretty frustrating — even putting aside my weak grammar knowledge, the lack of furigana made looking up words a struggle. Eventually I shelved it to read when it was more in my strike zone (which only took about a year and a half).
A mad rush through Bunpro
Oct 29th, 2018
I had actually tried Bunpro before (according to my email archive, I signed up in January), but for whatever reason I didn’t end up sticking with it. The team released a ton of content and feature updates to mark the site’s first birthday, and I decided to give it another try.
Starting Bunpro marked a major turning point. I quickly recognized it as the key to getting out of my years-long rut and decided on a whirlwind pace to complete the course all the way through N3 before the December JLPT exam, which I had for some reason signed up for.
My lesson schedule was:
- N5: all 100+ grammar points on the first day, as “review” from Textfugu
- N4: 10 grammar points per day
- N3: 10 grammar points per day
- (JLPT exam, short break)
- N2: 3 grammar points per day
- N1: 1 grammar point per day
My one exception to this was the 敬語 section of N4, which I ignored until much later.
In addition to my new lessons, I would also be making sure to reach zero reviews due at least once per day. I quickly discovered that the Textfugu content I had (eventually) completed didn’t even cover all of N5, and that a lot of the material I was learning was brand new to me. I struggled through some difficult days (even stealing time here and there while on a family vacation to Ireland) to make sure that I stayed consistent with my reviews. It was a bitter pill to swallow but I’m glad I did it.
My (real) first manga
Nov 3rd, 2018
This was the day I finished Yotsubato vol. 1 and felt — for the first time in years — as if all my struggle had been worth it. I was still only a few days into Bunpro at this point, but even so I was more than equipped (from a grammar perspective) to handle everything it threw at me. Reading was slow and methodical, but it was starting to happen.
First attempt at JLPT
In the weeks leading up to the exam, I pushed hard on Wanikani in order to reach level 35, which was supposed to give 95% coverage of all N3 kanji. In retrospect this was probably not as helpful as I expected, since knowing a kanji doesn’t necessarily translate into knowledge of the vocabulary that it appears in.
Dec 2nd, 2018
I arrived at MIT to take the N3, sat down in the waiting room, and finished up my last batch of N3 grammar lessons on Bunpro just before the exam doors opened. Needless to say, I was less prepared than I should have been, but at the very least the impending deadline had lit a fire under me. Pass or fail, I would be in a much better place going forward.
The exam was mentally exhausting and I remember being pretty convinced that I wouldn’t pass. With the benefit of hindsight, I can easily recognize the fact that I was still largely in “decoding” mode rather than “understanding” mode. For listening in particular, I was mainly listening for key words rather than trying to understand the sentence as a sentence. Listening would become an area of particular focus in the near future.
With my JLPT sprint over, I reduced my rate of new lessons on Bunpro and continued at a relaxed pace pace for the remainder of my subscription term, which allowed me to finish all of their available material in just under a year. It was around this time that I developed my personal strategy of skipping anything I didn’t understand, with the idea that reading more total sentences would give me a higher chance of encountering the grammar I was studying.
Dec 3rd, 2018
The very next day after the exam, fully convinced that I needed a lot more exposure if I was going to turn my new knowledge into real language mastery, I sat down to embark on a playthrough of Persona Q2, New Cinema Labyrinth. I livestreamed myself playing through the entire game (aside from some sidequests) and translating as I went. Looking back at the videos now, it’s astonishing how much trouble I had with things that are instantly obvious to me now.
With N3 under my belt, I found it very easy to jump into pretty much any manga I found interesting. There were still gaps of course, and I was frequently making mistakes and getting lost. But ultimately I was having a good time and learning more every day. This was the point I had been trying to reach since day 1, and all that was left was to put in the time.
First JLPT results
Jan 22nd, 2019
I got my results back and, as expected, I had not passed. With a total score of 73/180, and a sectionwise score of 28/60 (language knowledge), 19/60 (reading), and 16/60 (listening), I was close to the minimum passing grade, but to be totally honest I’m sure some of those points came from bubbles that I filled in at random. In retrospect this was a stupid move — I stood to lose nothing by failing and would have gained nothing by an unearned pass.
March 1st, 2019
Coincidentally, exactly three years from day 1, I posted my first reading report. I would write one of these every single week for the foreseeable future. I often look back at them to get perspective on how far I’ve come.
Wanikani level 60
Dec 4th, 2019
Reaching the max level on Wanikani felt like the end of an era. It had been with me since almost the very beginning of my journey, and now I had finally conquered it. I polished off the rest of the lessons over the next couple of weeks and wrote my customary retrospective post on the forums. As my reviews slowly tapered off, it would gradually become less and less of an important part of my study routine.
Jan 29th, 2020
With reading going pretty smoothly, I wanted to get my listening skills up to parity. I watched all of ヒナまつり in one week and reviewed it with an audio flashcards method that I wrote about here.
No longer studying
Mar 22nd, 2020
Four years to the day since I started Wanikani, I decided to largely put SRS behind me. The diminishing returns from Wanikani’s high levels had started to set in, and my daily vocabulary drills in Anki, which had always seemed like a necessary evil, no longer felt necessary. I would continue using Subs2SRS to improve my listening, though never as a daily habit.
At this time Bunpro’s N1 material was still under development, and I still had a nearly-untouched copy of 新完全マスター N1文法 sitting on my bookshelf. It was my intention at the time to get around to it eventually, but at the time of writing I still haven’t gotten around to it (and I don’t feel the need to do so either).
The long reading grind
For the next two years, I read for a bare minimum of 30 minutes every day. My comprehension leveled off at some point and I started noticing that I was almost never coming across sentences that I didn’t understand. My reading speed continued to climb, but slowly. I wrote year-in-review posts for 2020 and 2021 and was encouraged by my improvement across the board.
This marks the end of my retrospective; all future events will be added as they happen.
Second attempt at JLPT
Dec 4th, 2022
After three years of missed opportunities (once because the test site filled up before I could get a seat; twice because the test was cancelled due to COVID), I finally had a chance to try the JLPT again. I decided to go for N1 this time, thinking it would be better to push myself instead of going for a (theoretically) easy N2 pass.
Leading up to the day of the test, I was feeling pretty nervous. I’ve always heard that each level is an exponential jump in difficulty from the one below it, and that N1 in particular is intensely difficult. I’ve even heard people give advice like “don’t try to read the whole passage, just skim for keywords” and “do the reading section before grammar/vocab or you won’t have time”.
As soon as I opened the test booklet, all of that anxiety drained away. The passages didn’t seem any more difficult than the material I read on a regular basis for fun, and I was able to fully read each one and answer all the questions with time to spare. The listening section was noticeably tougher, but there were a few questions in particular that actually got a barely-stifled laugh out of me because of how obvious they were.
I left the exam site feeling pretty positive about my chances at passing, and surprised at the fact that I didn’t feel any kind of exhaustion or burnout from three straight hours of testing. Later that night, debriefing with some friends who had taken the same test, it was pretty evident that I had made some mistakes, but also that I had gotten the right answer on some questions that were designed to be tricky. Results wouldn’t be out until early in the next year, so all that was left to do was wait.
Second JLPT results
Jan 22nd, 2023
The results finally came out, and I hadn’t passed. It was a bit of a disappointment, but on the whole I wasn’t too bothered. Passing or failing wouldn’t have any impact on my immediate plans and considering I actually enjoyed the actual test-taking process, I had already been planning to take it again the following year.
My main takeaway was that I didn’t plan to change my approach whatsoever. When I do eventually pass, I want to have achieved it through natural input, not grinding textbooks.