Pronunciation ぢ じ ず づ

Howdy Crabigators,

I just came across 鼻血 in vocab, and the preferred spelling was はなぢ. I realized that I’d never really come across that before, so I looked up how to pronounce it, and EVERYWHERE said it’s voiced exactly the same as じ. Ok. Makes sense.

Where things took a turn, is that most of these sources also said that ず and づ have the exact same voicing. I have always said zu and dzu, respectively, and I SWEAR I hear づ as dzu in use. Was I imagining dzu all these years? Is it not Suzuki ga sudzuku, but Suzuki ga suzuku?


Short answer: Yes. Possibly springing from a tendency of fansubbers to render づ as “dzu” when they’re doing song lyrics in romaji.

Long answer: It’s… kinda complicated.

(Sorta related fun fact: “Kudzu” and “adzuki” are both Japanese loanwords in English, but neither word has ever been spelt with a づ in Japanese - they’re くず and あずき.)


So when I say ず my tongue is in the position to make a sssssssss sound

When I say づ I do the exact same thing, except the middle of my tongue taps or almost taps the roof of my mouth to give a hint of a d sound.

Are either of those unnatural? To my ear it sounds right, and any of my Japanese teachers were never very critical of pronunciation outside of hammering ら り る れ ろ (I’ve been trying to get into pitch accent, and it’s maddening that none of my teachers brought that up)

1 Like

What would sudzuku or suzuku be?


That would be me making typos…



Ah, gotcha.

But anyway, as noted by the yotsugana link Belthazar posted, there’s no one answer to this. Some distinguish them. Some don’t. And unless a particular Japanese person you’re talking to is well-versed in linguistics, they may not even understand what you’re asking about.


This native Japanese guy explains a bit about Yotsugana. Hope they could help a little.

Related video that goes into the pronunciations:


i am commenting to let you know i have no answer but have the same question

1 Like

Both ず and づ are sometimes pronounced dzu by some speakers. I think. Feel. I’m fairly confident. This is bound to confuse some people.

1 Like

Mei-san in the videos above confirmed these by the two different pronunciations used.

Usage of [dz] for:

  • At the beginning of words
  • After ん
  • After a small っ

Usage of [z]

  • In the middle of words
  • At the end of words

After having listened to about sixty variations on 残念, 随分 and 全然 on Forvo, I have come to the conclusion that I spend too much time on this kind of minutiae and my ears no longer know what is dz and what is z. This is fine, I guess.

Thanks for the breakdown of the video.

1 Like

I’ve come to the point that if native speakers could understand what you mean, that’s all fine since the difference in these cases are just very slight tongue position differences (voiced alveolar fricative vs. voiced alveolar affricate). Just like how English has various pronunciations and accents, it’s usually a matter of regional differences or personal preference.

Unlike the ones with the same kana but different pitch with different meanings. I’m still working on those with the help of my Sensei, who usually doesn’t let them pass, lol.


The only way to answer this is to collect recordings you thought might be different, then do some analysis digitally? (Or ask someone to help.)

Personal pronunciation is a different question. Perhaps you need to ask a skillful listener? Perhaps a polyglot or a linguist, rather than a native.

About Kana usage, I found that they may change, and sometimes there may be multiple variants accepted. Sounds may change as well. It is also possible that sometimes variants aren’t accepted for a commonly typed word, but natives might not hear the difference in practice. Kana aren’t really sounds, but things like manga may follow the sound more closely.

1 Like

I… dug more, because I still hear dzu and zu varying somewhat freely, both word-initial and inside the word. Apparently, a study has already been conducted by a linguist on this topic, and they seem to agree, although the rules of thumb shared earlier do correlate to real-world usage, most of the time, due to other factors.

Contrary to the traditional linguistic account, the variation is not a positionally conditioned, hence categorical, allophonic variation. Although the effect of linguistic position exists to some extent, its influence is secondary compared to the influence of local temporal characteristics of speech called time allotted for consonant articulation (TACA). The overall prediction rate of the manner of /z/ articulation by means of TACA was 74%, and, when coupled with information on linguistic position, the prediction rate was as high as 80%.

The answer still seems to be “it’s complicated”, but dzu does not seem to be a figment of imagination, at least.


Great link! The link to “Development of allophonic realization until adolescence: A production study of
the affricate-fricative variation of /z/ among Japanese children” has oscillograms, spectrograms, and other graphs that will make me procrastinate for my Japanese reading class tonight. :see_no_evil: