Why do kanji have so many homophones?


Can anyone with a strong knowledge of Japanese/Kanji weigh in on this?

I just can’t understand why there are so many kanji with the same pronunciation.

There are just tons of kanji that all have the same ”こう”、”たい” etc. pronunciations.

Mathematically, this just doesn’t make sense.

There are 46 basic kana. Even if we just limit kanji to having two basic syllables, to get the number of unique kanji sounds you could make, you would take 46 P 2.

This is 2070 different pronunciations. That’s enough to cover pretty much all of the necessary words for basic communication.

But kanji can have up to three syllables. If we include those, we end up 91,080 different pronunciations. That’s enough to cover an entire language.

But there’s actually more than 46 sounds in Japanese. Counting all of the advanced/extended sounds, you get 114 different syllables.

So, 114 P 2 gives us 12,882.

114 P 3 is 1,442,784.

Over a million possible unique sounds that each kanji could have.

So, if I only know about 100 kanji, and there are over 1 million possible pronunciations for kanji, why the hell have I just learned about the 10th kanji that’s just pronounced “こう”?


I’ll leave this here.

Most simple explanation. Kanji are ideas, not vocabulary. But that’s not true but it is, but I’m not going to rant on this. :sweat_smile:


Well, there’s no one person who decided to make Japanese this way, it just happened over time. All languages are illogical, you might as well ask why English pronuncation is the mess that it is, or why romance and germanic languages have gendered nouns or whatever.


My problem is actually the opposite. Not why each kanji has so many pronunciations, but why so many kanji have the same pronunciation as other kanji.


I know. You’d just think somewhere along the line someone would have though, “Hey. This one pronunciation can mean 20 different things. That’s kind of confusing. Maybe we should change that.”

I know every language has homophones, but the amount in Japanese seems way more excessive than any other language I’ve come across.

And a lot of other facets about Japanese grammar seem to be very deliberate and logical. All of the homophones just seem like a big curve ball for what is otherwise a very structured and logical language.


Imported from China

Chinese is tonal language

One spelling to say many things. Different tones to differentiate.


Bingo. Blame Chinese.


Exactly. And then Japanese is like let’s adopt the new readings when power changed. There’s no way it’ll change a third time.



Removing tones definitely makes it worse, but that shows that tons of homophones still exist in Chinese characters, even counting for different tones. I didn’t realize Chinese had so many homophones, too, but that just makes it more confusing.

Like, both languages have undergone lots of changes, where people in power would make deliberate choices on rules and pronunciations. As your first link showed, with each dynasty change, there were big changes to the language. Each of those changes would be a chance to address some of the problems.

And again when things were brought to Japan, that would be another chance to change things up. Japanese, especially, has had a lot of very deliberate and big changes made to the language over time. Introducing hiragana. Introducing katakana. Simplifying Chinese pronunciations.


Gah. It’s just frustrating.


Like, with a lot of the grammar for Japanese, it’s confusing to learn as a Westerner, but you can see that there is a lot of structure, and that it will probably make a lot of sense once you actually get a hold on things.

It’s a goal to work towards.

With the homophones, it’s the opposite. It’s pretty clear that it doesn’t really make sense, and it probably won’t ever make sense. It’s just something you’ll have to remember by drilling and brute force, and even then you’ll probably make simple mistakes for years to come.

It’s pretty demoralizing.


First of all… Two kanji that share an onyomi are not homophones. That’s the wrong word, unless they are also words in that standalone form.

There are lots of homophones in Japanese, but having lots of kanji that are pronounced こう does not automatically make that the case.


I know a lot of them aren’t exactly homophones, I just don’t want to keep typing “kanji that have the same pronunciation”. I’m well aware the vocab pronunciations are often different, or that you need okurigana for the words to make sense. There’s still a ton of homophones, even accounting for those, though.


And that’s where the loss of tones comes in. Because you can have 60 different こう kanji with way fewer homophones as a result if you add the tones in. Because in Chinese, generally you always add two kanji together to make words, so as long as you are adding different こう’s with different other ones, you don’t have any homophones. Chinese does have homophones, so it’s not like they completely avoided it, but I don’t see why it’s hard to imagine Chinese originating in the way it did.

Take something like こうせい, which there are a bunch of in Japanese. It’s easy to imagine how if you had 4 tones, you could actually combine the こう’s and the せい’s in such a way that you effectively only had a handful of homophones instead of 20 or whatever it turns out to be in Japanese.


There is a common misconception across language learning which holds back a lot of people, myself included. Just like with sheet music, written language is a representation of spoken tongue. Just like theory tends to come from practice and then informs education so people can put it back into practice.
This logic applied to Kanji would show that your problem is not as such with Kanji, but more with Japanese for having as many homophones :sweat_smile:.


I wonder if the Chinese words based on kanji were widely integrated in the language from the start. The kanji were imported as holy texts in the beginning. Not exactly the place to complain that the readings are just too confusing, and the monks had the text to check the meaning, not sure if they actually spoke like that.

Then the imported words were used in a court setting, where people are proud to make things deliberately harder to impress others. And after wider usage it’s not that easy to change things after 500 years of reading it that way, and maybe seen as impolite to your ancestors or you would look like an uneducated fool if you switched the reading from the traditional form. It’s like insisting to count oneteen, twoteen because it’s more regular.


Long image, but it always makes me chuckle.



While completely inaccurate, still funny. Fucking rendaku…


A syllabary can only have so many sounds. Japanese only has a little over 100 of its own, I think…


It doesn’t make much sense to randomize the sound inventory and assign new sounds to each one of the kanji, since they were brought over as loanwords. They’re not going to go, “oh this is the tenth こう kanji, let’s make this one じあ instead”.

In Japanese バス can be “bath” or “bus”, because both things got simplified into the same Japanese sounds, but that doesn’t mean that Japanese should assign completely new sounds for one of them so they’re no longer homophones.


Wait, are you trying to suggest that’s not what really happened?