How dakuten and handakuten change a kana

Tenten, maru, and what they are called

Japanese has one name for each mark and another for the sound it writes. The two strokes are 濁点だくてん (dakuten, "muddy mark"), colloquially てんてん (tenten, "dots"). The circle is 半はん濁点だくてん (handakuten, "half-muddy mark"), or まる (maru, "circle"). The sounds have names too: 濁音だくおん (dakuon) for the が, ざ, だ and ば rows, 半はん濁音だくおん (handakuon) for the ぱ row, and 清音せいおん (seion, "clear sounds") for plain kana. A dakuten is the mark; a dakuon is what you say.

You will meet the marks constantly. Yominews, Yomimaru's news section, had 156 graded news articles in September 2026, or 289 texts once each level version is counted. In those texts about one printed kana in eight carries a mark, and で and が alone make up over half of the marked ones. Both are also grammar particles.

Why the は row also takes the handakuten

The か, さ and た rows take one mark each. The は row takes two, and ば (ba) is not the voiced ha the way が is the voiced か. The は row is thought to have begun as a p sound. It softened to something like f, then to h, and ふ still keeps a soft f. ば stayed its voiced partner, so writing p again took a second mark: the small circle, first used by Portuguese Jesuits in the late 16th century (the history of both marks is on Wikipedia).

When do you write ぢ and づ instead of じ and ず?

Only in two cases, and your ear cannot tell you which words they are. In standard Japanese, ぢ sounds like じ (ji) and づ like ず (zu); a few dialects in Shikoku and Kyushu still keep them apart. Japan's spelling rules, Gendai kanazukai (現代げんだい仮名かな遣つかい, a 1986 Cabinet notice), write じ and ず by default and keep ぢ and づ for two cases:

• When the sound repeats: ちぢむ (chijimu, to shrink), つづく (tsuzuku, to continue).
• When two words join and the second began with ち or つ: はなぢ (hanaji, nosebleed) is はな, nose, plus ち, blood. みかづき (mikazuki, crescent moon) ends in つき, moon.

The second case is rendaku (連濁れんだく), the voicing that often happens where two words join, as in て plus かみ giving てがみ (tegami, letter). A word no longer felt as two, like せかいじゅう (世界中せかいじゅう, sekaijuu, around the world), is spelled with じ, although ぢ is also allowed.

I only found out ぢ existed after a year of studying Japanese. Since then I have asked friends at my Japanese language school to name one word written with ぢ, and so far none of them has managed it. The counts below show how rarely it turns up.

To type them with a Japanese input method, enter di for ぢ and du for づ; ji and zu give じ and ず. None of the 13 ぢ words is tagged N5, N4 or N3, which is why the list below opens with づ words.

ヴ, yōon, and marks outside the chart

One more kana takes a dakuten: ウ, which becomes ヴ (vu) for the v in loanwords. Japan's 1991 guidelines for loanwords list ヴァ, ヴィ, ヴ, ヴェ and ヴォ, then note that such words can generally be written with バ, ビ, ブ, ベ and ボ instead. Both spellings are in use, ヴァイオリン and バイオリン, and the dictionary files the word under バイオリン. Most speakers say ヴ as b anyway. A hiragana ゔ exists too, in Unicode since 1993, but it turns up only occasionally, in names for example. In the Yominews articles, ヴ appears 38 times and ゔ never.

The marked kana also combine with a small ゃ, ゅ or ょ, just as plain kana do: ぎゃ (gya), じゃ (ja), びゃ (bya), ぴゃ (pya). These combinations are the yōon (拗音ようおん), and Yomimaru's free kana course teaches them right after the marked kana.

In manga you will sometimes see a dakuten on a vowel, like あ゛, for a strained, throaty cry. It is not a real voiced sound.

What does the small tsu (っ) do?

A small tsu, っ (ッ in katakana), is not read tsu. It is the sokuon (促音そくおん): it holds the next consonant for one beat before releasing it. Compare きて (来きて, kite, come) with きって (切手きって, kitte, postage stamp): in kana, the small っ is the only difference. In romaji the consonant after it is written twice, except that Hepburn writes a doubled ch as tch: まっちゃ is matcha. At the end of an exclamation, as in あっ or きゃっ, it marks a sudden cut-off instead, and romaji has no standard way to write it.

It is common: one printed kana in 62 in the Yominews articles is a small っ or ッ. With a Japanese input method, type the next consonant twice (kitte gives きって). For a small っ on its own, type xtu; ltu also works in most input methods, including Google Japanese Input. The katakana side, and why ッ is not ツ, is on the page about telling シ, ツ, ソ and ン apart.