Japanese Rules-of-Thumb/Heuristics

Hi all,

I wrote a blog article which tries to gather all the “rules-of-thumb” or heuristics I have found useful in learning Japanese. The reason I wrote this is because none of these are written in standard textbooks, and knowing them really accelerated by learning.

I would be very grateful for any feedback, e.g. things you disagree with on the article, or things you would suggest adding. A lot of it is based on things scattered across various Tofugu-pages, but there is some stuff which isn’t at all related to the great Grabigator.



This one seems like it needs to be rewritten to me.

It is rare to see a Chinese (on’yomi) reading that is more than one syllable, so if you know a reading with many syllabus「みなみ」 (南) you can be fairly sure it is the Japanese (kun’yomi) reading.

Two mora onyomi readings are very common, so maybe you meant “more than two” above? But then you would need to change “It is rare to see” to “You will never see.”

Onyomi are always either one mora or two.

Or maybe the somewhat vague word “syllable” is what is responsible for it.


Thanks, you’re right! Maybe Syllable is too ill-defined in Japanese. I have corrected it to 1-2 mora.

I didn’t thoroughly read it but I looked through it and it seems good! A lot of the information you’ve collected here are things that I’ve found useful and try to pass on to other learners as well. I did happen to notice a couple typos:

“I took way that I am bad at learning languages”
“心底 is そんそこ

I’ll look through it in more detail later if I have some time.

1 Like

Heuristic : Most of the time, if the first kanji is Japanese reading (kun-yomi), the second kanji will not be rendaku’d.

This one’s really handy; I hadn’t made that connection and the かわぞこ vs. しんそこ thing really bothered me.

Phonetic-Semantic Decomposition is really useful; I’ve gotten into the habit of dumping new kanji into Wiktionary which usually explains which kanji are phonetic-semantic, or if it’s an ideogrammatic compound. I frequently just skip Wanikani’s phonetic mnemonic, because it’s not necessary if I recognize the phonetic component.

I believe the phonetic component is often chosen to carry a slight semantic meaning, too, which can be useful to know, too.

1 Like

Thank for making this, looks very interesting! Just a few random comments:

“Phonetic-Semantic Decomposition”, wow, that’s a terrifying word, but all it really means is that there are kanji which you can predict the Chinese (on’yomi) (“phonetic”) reading just by looking at the kanji’s components (radicals)! This works for about 1 in 5 Kanji (22.06% of Kanji to be precise, see here for the statistics and science).

I also discovered that site before and found it very interesting, but after working through all the kanji I think the 22% is way too conservative. The readings are just as messy as Japanese itself. For example 方 didn’t even make it on that page, although it’s a very good indicator in my opinion.

It’s problem that it can be read either ほう or ぼう, which are the readings from the Han and Wu dynasties. A better way would be to look at all valid readings and the occasional sound shift (like しゅう → ちゅう or something) to make a judgment, at a random guess the number should rise to maybe 50% (from the ~66% of the kanji that are phonetic compounds in WK).

This pattern works because historically in Chinese, the right-hand side radical was related to the pronunciation (“phonetics”) and the left-hand side was related to the meaning (“semantics”).

As a heuristic this is OK, but I don’t particularly like “left-hand/right-hand side”, there are some radicals that go right like 刂 or 阝, and in some rare compounds the phonetic component is not what it seems (like in 視 which has 示 --し), some components like 心 can be all over the place, like 志, … Still, it’s mostly left-right.

1 Like

Thanks everyone! I’ve fixed the typos mentioned.

at a random guess the number should rise to maybe 50% (from the ~66% of the kanji that are phonetic compounds in WK).

If you consider “imperfect” phonetic decompositions, i.e. ones where there is a rough trend, then that website (namakajiri.net article) does discuss this:

456 Jōyō means 21.29% are perfectly phonetic, and 1411/6394 = 22.06% of our larger set (the yellow, >imperfect part goes up to 34.46% and 49.69%

So your guess of 50% seems quite plausible. But if there is a sound shift, then you are going to have to remember an exception anyway, so I’m not sure the “imperfect” phonetic decompositions are quite as useful. Regardless, I have edited the article to explain this, thanks!

Good point, I’ve also updated it to make it clearer that left-right is not the only place they can be arranged.

It’s less useful for sure, but it is often better than nothing, and I find it easier to remember a transformation than starting from scratch.

I just have trouble calling something “imperfect” that is perfectly regular (there are lots of ほう/ぼう, ふ/ぶ, もう/ぼう and so on). The Japanese way of importing just created a moving target.

Maybe I’m just saying that it helps to learn about the different kinds of on-yomi, the multiple ways of reading a kanji make much more sense afterwards.

1 Like

I can testify they are most definitely useful.

Also, shouldn’t those be semantic-phonetic compositions, not decompositions?

1 Like

Yeah I guess I use the imperfect ones to some extent too.

(Edited it to “compositions” rather that decompositions)

1 Like

He used the word correctly in that しょう and せい are indeed one syllable, but two morae. Of course the word is “vague” in the sense that it’s often misused (kinda how you often don’t quite know what people mean when they say “factoid” or “peruse”).
However, there are two syllable on’yomi readings like がく which are clearly two syllables so his statement is not correct with syllable instead of mora either.
(These readings still derive from a single syllable of Chinese, in this case presumably something like the Middle Chinese /ɦˠʌk̚/, with the added u to make it into Japanese)

You don’t see the concept of syllable much in discussions of Japanese because it’s usually not a helpful concept. Japanese native speakers find morae more natural than syllables; it fits their understanding of their own language better.


When some kana’s come out
But a vowel’s en route
That’s a morae

1 Like

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.