Rendaku rant

Rendaku is driving me crazy to oblivion!! I know there are rules, I’ve tried to understand the Tofugu article, and most of the time I can figure them out, but I still can’t figure out why, for example, 点数(てんすう)doesn’t rendaku but 人数(にんずう)does… which case/rule would this fall into? Or do I just have to memorize the hard way? Argggg…

6 Likes

Exposure sorts things out over time. I guess if you look at that article, 点数 is the normal one (a Chinese origin word that doesn’t have rendaku), which makes 人数 the exception. I suppose you could think of it like that if it helps.

But eventually it’ll just get to a point where てんずう and にんすう just don’t sound right, because てんすう and にんずう are common words you’ll hear over and over and just get used to.

The rules of thumb might be good for helping you guess better than chance when encountering new words, which is nice, but for common words, you’ll just have them drilled into your head so much you won’t have a choice but to remember the rendaku status.

14 Likes

There’s this I remember someone mentioning being nice (probably, can’t remember), though it’s quite old, probably worth a check

5 Likes

I found a reddit post where somebody asked this question, and one of the responses was interesting (if speculative). Apparently there was a period in the history of the language, in Early Middle Japanese where you couldn’t have an ん next to an unvoiced consonant in the middle of a word, so the those consonants got voiced (thus 読んで, 呼んで, 死んで ending with で instead of て). So a ‘*nsuu’ word that was old enough that it predated that period would end up with the ‘*nzuu’ sound, whereas a word that was coined later when that rule was no longer operative would be OK to be ‘*nsuu’.

As the poster says, though, you’d need to actually check whether these two words had datings that fit with the theory before it becomes more than interesting speculation.

As a practical matter, though, I agree with @Leebo . I’ve never studied rendaku rules; I think that like kanji readings, (1) they can be useful but not 100% reliable guides if you need to guess the reading of an unknown word; (2) if you ignore them and just learn words then eventually you’ll find that you’ve subconsciously acquired good enough pattern-matching skills for making better-than-random guesses at readings of unknown words anyway.

8 Likes

This is pretty much the strategy I’ve come to use for recalling readings that have given me trouble thanks to anything like rendaku or one or more of the kanji having similar but still different readings to pick from (like じん vs. にん for 人) Especially after I’ve reviewed a word enough times, the correct reading just tends to “sound right” in my head.

I’ve been tempted to try studying the rules of rendaku to make things easier, but I’m not really sure if it’s worth the effort when you’re still going to run into exceptions that you just have to memorize anyway.

3 Likes

It’s an instance I find remembering Romaji makes more distinction than Kana, so I made a mnemonic based on English. Otherwise, seeing Romaji, than imagining later, works too.

Sometimes I hear difference between s and z, sometimes I don’t. My native language doesn’t make this distinction.

I don’t remember Rendaku rules, and I still don’t guess that well. With IME, I just type twice.

to this day I dont even try to understand them, just pure memorization for those.

2 Likes

I still have to repeat 老人ホーム to myself in my head to figure out which is which for 浪人 and 老人 :smiling_face_with_tear: It does work every time though.

3 Likes

Those two’s vocal distinction is very clear for me. Also, I know both words. So, no problem.

Otherwise, without knowing the vocabularies first, 者 for しゃ, じゃ, もの is very problematic.

人数(にんずう)probably needs to be distinguishable vocally from 日数 ( にっすう) , and the other word doesn’t have such a noisy neighbour that pushes it away. I think the need for words to be apart from similarly sounding words is what often drives the rendaku. So, there’s no rule to help with learning, it’s only once you know all the words and can recall the similar-sounding ones from the initial one, you can guess at reasons.

Are there even rules that are absolute? I’ve never studied any rendaku rules, but I don’t think I’ve picked up on anything completely reliable anyways.

If you are a native English speaker, one question I have for you is have you ever misread Come as Coam?

I mean, we have this whole silent e pattern right? Like home, bone, lone… but there are a ton of exceptions. So how do you know without any problems? Well, you’ve heard it so many times and know the word. Same for japanese. Once you know the word it’s not actually an issue anymore.

5 Likes

For those interested in linguistics, here’s an interesting bit about some rendaku stemming from Chinese from an old post I made from a now-deleted reddit account.

Chinese characters in Japanese is an interesting case. The onyomi (Chinese-originated reading) that is polysyllabic is usually derived from the ‘checked tone’ (入声) of Middle Chinese (spoken around 500 AD to 1500 AD). For example, 六 read like ‘lok’ in Middle Chinese with a short ‘k’ sound at the end. When the Japanese adapted the reading at that time, a ‘u’ sound was added to the end to make it sound like ‘loku’ because Japanese doesn’t allow words to end with a consonant that is not ‘n’.

Modern mandarin abandoned the checked tone mostly by dropping the ending consonant so that all characters are monosyllabic in mandarin today. However the checked tone is preserved in many dialects like Cantonese, Hokkien and Wu. So polysyllabic characters still exist in these dialects.

When I began to learn Japanese I was baffled by the fact that 六 is pronounced ‘loku’ on its own but ‘lo-pon’ in ’六本’ where the ‘ku’ sound is completely dropped. Then I learned the checked tone in Middle Chinese and I realized that it makes perfect sense because it is just so natural to drop the short ‘k’ sound in Middle Chinese when it is followed by another consonant, which also happens a lot in English. Since then it became a fun game to spot these polysyllabic onyomi in Japanese and validate them in the Chinese dialect (Wu) I natively speak and surprisingly it never fails.

https://www.reddit.com/r/ChineseLanguage/comments/6cv5md/comment/dhybumx/?utm_source=share&utm_medium=web2x&context=3

4 Likes

When you read this sentence you naturally assume “read” was past or present tense almost instantly. It just takes practice and I wouldn’t get too butthurt about it as it’s gonna come up in listening, reading, conversation, etc. Best not to get caught in weeds if you are like me and just are a more kinetic learner than a textbook one in some areas.

I’m normally a “by the rules” learner and like to pick apart the reasons for why things are the way they are, but I’m always willing to make exceptions for things where studying any written rules won’t really help you that much in practice. This seems to be one of those things.

2 Likes

Trouble with written rules is that in Japanese, All Rules Have Exceptions, Including This One. The Tofugu article that the OP linked to helps, but yeah.

Coming from English, a language that seems to love breaking its own rules, yeah I get that every rule will inevitably have exceptions, lol. It’s just that Japanese generally feels more regular in its structure than English (at least from what I’ve seen in my grammar studies so far, I could be wrong) so exceptions or irregularities seem to stick out more.

I’m sure there is A reason, but I don’t think knowing the how or the why will make it any easier or less time consuming as opposed to just memorizing the word in isolation.

1 Like

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