Six Months of Japanese — Progress Update

Previous Posts:

1. One Month of Japanese
2. Two Months of Japanese
3. Three Months of Japanese
4. Four Months of Japanese
5. Five Months of Japanese

(Note that I am counting months of study, not calendar months. I started studying on Dec. 14, 2024.)

Total Time Studied: 405 hours

Total Hours of Extensive Listening and Reading: 72 hours

Average Daily Study Time: 2.3 hours (up from 1.89 hours last month)

Total Vocabulary: 9,100 words

A quick note about my vocabulary estimate: I arrive at this number by counting the number of new words I've learned each day and entering that number into my spreadsheet, then totaling that number over time. There are several unavoidable inaccuracies in this number, including the following:

• Words I have learned, but since forgotten
• Words I have learned, but not counted (e.g. I learned them via extensive input)
• Words that are immediately transparent to me based on words I've already learned, but I haven't officially "learned"
• It doesn't account for "degrees of knowing," i.e. words I have a vague understanding of are counted the same as words I'm deeply familiar with
• Inherent difficulties in defining what counts as a separate word

I do not believe my vocabulary count could be realistically off by, like, an order of magnitude (which is why I consider it a useful number), but my gut feeling is that the "true" number could be plus or minus several hundred.

Link to Spreadsheet

Quick Disclaimer:
There was some confusion last time. I am not Chinese. I do speak Chinese, and I learned the language to a level sufficient for reading some fiction written for young adults (with a vocabulary of about 20k words), but I am not ethnically Chinese, and I did not grow up speaking Chinese.

My Study Routine:
I often, but not always, get some reading in immediately after waking. Typically, this will either be 1-3 news articles, or part of the novel I am reading. (I'm still working my way through ライオンと魔女.) That takes me 0.5-2 hours. I review my old flashcards in Anki shortly thereafter. On average, this amounts to 400-500 cards and takes me 0.75-1 hours. I review new flashcards shortly before bed. New flashcards number exactly 80, and take me on average 0.5 hours to get through.

Any additional studying I do is optional. Examples include reading Wikipedia articles, watching informative videos on Youtube, and watching news broadcasts. I am not working on developing speaking or writing capability.

Improvements in Listening Comprehension:

There's a sensation, and I'm sure many of you will know exactly what I mean when I describe it: You're listening to a stream of speech, and your mental processing speed (ability to match sounds to words, words to meanings, and collections of words to more complex meanings) is just a little bit too slow. You frequently catch phrases of 4-6 words, and much of the rest of the time, the speech is tickling your brain. Like, you can somehow feel that the words that are being said (that you are completely failing to parse) aren't unknown. If only your processing speed were a bit faster, you'd be able to understand dramatically more.

That's where I'm at right now.

I experience the sensation I described above very strongly with news broadcasts about politics and international affairs. (I'm not specifically limited to such narrow domains anymore—see previous updates.)

I've started to understand at least some of Dogen's skits, which feels fucking weird.

I've started watching videos like this one, this one, and this one, to train my listening comprehension. If I need to harvest vocabulary from a Youtube video, I use this transcript generator.

My listening comprehension seems to be advancing significantly faster than it did with Chinese. I'm…not sure why. Theories:

• Focusing my efforts on limited domains has made it easier for my brain to latch on to familiar vocabulary
• The large number of cognates from Chinese is helping (but how can that be, when all of the cognates sound completely different?)
• I acquire listening comprehension in new languages faster than before, simply because I've already done it with five other languages
• It's not that Japanese is particularly easy. Rather, Chinese is particularly hard (Chinese lacks audible word boundary cues, i.e. past tense suffixes and other word-final morphemes)

I consider cracking listening comprehension to be extremely high priority, for the following reasons:

• Japanese people speak much, much faster than I can read, which means listening to audio is always going to be more efficient immersion (based on words per minute)
• Good listening comprehension enables me to study while doing other things, e.g. washing the dishes
• My experience with Chinese taught me that having excellent reading comprehension and terrible listening comprehension is kind of a miserable experience, and I don't want to repeat it.

Improvements in Reading Comprehension:

News articles are increasingly easy going for me, and Wikipedia articles are very approachable now. I am no longer limited to the extremely limited domains I originally chose to saturate my vocabulary in. For example, a few days ago, I read this article about the power consumption of LLMs, and this article about ongoing demonstrations in Serbia. Neither was particularly challenging—I did make use of Yomitan, but not a huge amount, to be honest.

I am able to handle drastically longer sentences than before. 6-7 clauses are almost never a problem for me in the novel I'm reading (though the clauses there are quite short). At least for relatively simple texts, I am much more likely to have a problem with an unknown grammar point than the simple length of the sentence. Particularly long sentences do still cause me problems in information-dense writing, like Wikipedia articles.

Also, Japanese's lengthy left-branching constructions cause me a lot fewer problems than before. I can still get befuddled if they are particularly long and complicated, but way less than before, and usually, if I give myself time, I can puzzle it out without falling back on machine translation.

I mentioned in a previous update that I had problems with unintentionally ignoring case particles and interpreting the argument immediately before a veb as the verb's subject—this no longer happens at all.

Improvements in Pronunciation:

I am starting to develop an intuition for which mora is accented on non-compound words. I've noticed that some morphemes seem to increase the probability that the accent will fall on a particular mora. Certain combinations of morphemes seem to also affect the probability of mora placement. In general, I've noticed that "no accent" appears to be the default, "accented on the first mora" is second-most common, and "accented somewhere in the middle of the word" (typically the third or penultimate mora) comes in a distant third. Words that are accented on the final mora (with downstep on the following case particle) seem to be exceptionally rare (yay!), EXCEPT for very common words (ugh) which are typically 1-3 morae in length.

I assume that pitch accent in Japanese is much like stress in English or tone in Norwegian, in that my accuracy in guessing the correct accent in unfamiliar words will gradually increase over time, but never exceed, say, 70% accuracy.

General Improvement:

I'm starting to notice "general utility" in my Japanese skills. The first example of this was when I started understanding the Japanese definitions in Yomitan. Recently, I've noticed an ability to navigate basic pop-up menus on Japanese websites. Then, I noticed that I've started to understand some of Dogen's skits. This is notable because these are uses of the language that I haven't explicitly studied for.

Ongoing Study Strategies:

"Reading" my novel (ライオンと魔女) involves reading line-by-line, with HEAVY use of Yomitan (sometimes it feels like every fifth word—I add every unknown word to my Anki deck), and learning about some new piece of grammar I didn't know about before roughly once per 1-3 paragraphs. I often have to consult machine translation to wrap my head around a particular sentence—this is usually because I know all of the words and all of the grammar in a sentence, but it doesn't "click" in my head until I have someone else tell me what it all means.

I had hoped to be at least halfway through my book by now, but as of now, I am still around Chapter 5 (out of 17). Mostly this is because I haven't truly prioritized this over other reading content. And that's mostly because working my way through the book is a slog, and I'm not always in the mood to bang my head against a wall of grammar for an hour. But Chapter 5 is already noticeably easier than Chapter 1. I'm getting there! In particular, I'm starting to be able to "read" (i.e. with Yomitan as a heavy crutch) longer and longer sections of my novel without needing to look up unfamiliar grammar. It's still slow going, but the improvement is noticeable. I'd say on average I can read maybe 3-4 sentences at a time without being puzzled by syntax, up from <1 sentence when I first started reading a little more than a month ago. I suspect improvement in this area will be very rapid in the coming months.

My study strategy is heavily influenced by an article I read several years ago, "Learning From General Word Lists Is Inefficient." (I strongly recommend reading it yourself; it was written with Chinese in mind, but the principle discussed is fully applicable to Japanese as well.) Based on the data presented in that article, I do not study from JLPT word lists. I do not use any pre-made Anki decks, ever. I harvest vocabulary only from sources I am likely to read or listen to.

I think I already mentioned this in a previous update? For complex numbers, rather than learning all base numbers in one go, and then learning a series of rules for how to combine them, and then practicing various combinations until I feel comfortable expressing numbers rapidly, I am instead memorizing random complex numbers (1884, 376, etc.) as they appear in my reading material. In my experience, this is an equally effective learning strategy in the long run.

I've started focusing a lot more heavily on developing my familiarity with complex numbers and number phrases (e.g. 1980年代、7倍 etc.) in the last week or so. I'm already starting to develop a good intuition for how pitch accent moves around in number phrases (e.g. 1920 vs 1920年 vs 1920年代) and can often (but not always) accurately guess correct pitch accent placement before I verify with an external source.

I've started harvesting vocabulary and grammar from BL erotica. I wouldn't have mentioned that, except that it turns out BDSM content is amazing for giving you a crash course in all kinds of formal language, informal language, and insults. So there's actually a very high volume of valuable stuff in there, and if you are into that sort of thing, I highly recommend taking advantage of it.

I haven't yet decided how I plan to handle Japanese names and surnames. Either I will memorize the readings of hundreds of surnames and hundreds of given names, or I will learn them as I encounter them. Probably I will do the first one, just to give myself a good base, but either approach has its merits.

Admitting You Were Right:

I got a lot of flack from commenters one or two months back for artificially capping my Anki review, and yeah. You were right. I raised my cap to ∞ and Anki is much less of a chore now. I still think it is important to make sure daily reviews don't climb to truly ridiculous heights (500 is already pushing it for me), so now I've been accomplishing that by aggressively removing cards that mature past roughly 1.2 months. I rely on intensive and extensive input for continued, "natural" SRS beyond that point.

This makes it incredibly important that I consume as much Japanese media as possible, and that the media I consume is as dense as possible (based both on words/minute and the richness and diversity of vocabulary used).

Study Methods I've Rejected:

• Apps and gamified learning (e.g. Duoling, Wanikani, etc.) — too low volume of new information, doesn't allow me to set my own pace
• Formal textbooks, courses, and classes — don't teach me what I want to know, when I want to know it. Tend to assume I don't have any prior experience with foreign languages. Teach me a lot of irrelevant (for my purposes) vocabulary
• Comprehensible Input (e.g. Comprehensible Japanese) — too little control over what I'm learning, how much, or how fast

Short-Term Goals:

• Finish reading ライオンと魔女 and begin reading my second novel within the next month.

Medium-Term Goals (achieve within 12 total months of study):

• Become comfortable with children's literature in Japanese
• Listen to at least one Japanese audiobook
• Listen to, and comprehend most of, a long-form news broadcast (15+ minutes) about familiar topics
• Watch at least one educational documentary about a topic of choice, and comprehend most of it
• Watch at least one movie

Long-Term Goals (achieve by the end of 24-36 total months of study)

• Read high literature in Japanese. By "high literature," I mean something on the level of Fifty Thousand Leagues Under the Sea. To be clear, I don't expect such reading to be easy. But I expect to have the understanding of vocabulary and grammar necessary to muddle through it at a reasonable pace.
• Read news articles about topics chosen at random with a high degree of comprehension
• Watch TV series and movies in Japanese without English subtitles, and understand most of what I hear
• Listen to audiobooks in a variety of genres, including nonfiction, historical fiction, science fiction, fantasy, romance, and erotica, with a high degree of comprehension.

Misc. Thoughts:

I know some of you wonder how I could possibly be sustainably acquiring 80 words per day. I've given some thought to it and come up with a number of possible explanations:

• At a speed of 80 words per day, other vocabulary frequently serves as SRS for kanji and vocabulary I've already learned. New words reinforce both meanings and readings of previously learned kanji. I suspect that learning 80 completely independent pieces of information would be much less sustainable compared to what I am actually doing, which is learning 80 new nodes in a vast, interconnected web of information.
• I've been learning languages as a hobby for over a decade now, and crunching vocabulary more or less the same way I'm doing now for around 5 years. I suspect that rapid memorization is itself a skill that improves with time, i.e. 80 words per day would not have been achievable for me 10 years ago.

I've noticed that the pitch accent of the recordings provided by my Yomitan setup do not always match the pitch accent notation in the dictionary. When I check natives' pronunciation on Forvo.com, it is virtually always the dictionary notation that was correct, and the Yomitan recording that was "wrong." Beginners, beware.

Much of my studying is done through the medium of Norwegian, rather than English. Helps keep my Norwegian fresh.

It blows my mind that y'all don't have a dictionary app like Pleco for Japanese. (It's a dictionary app that the Chinese learning community uses, and as far as I can tell, it completely blows everything Japanese learners have out of the water.)

I particularly enjoy having a cup of green tea or hojicha while I study. My little piece of Japan. I do miss it there. (I am in the UK now.)

I think that's everything for now. I'm now a quarter of the way through my originally planned 24 months of study! That feels wild. Looking forward to seeing what I can accomplish in the next six months.