Hello I'm trying to understand something about kanji readings. As far as I understand each Kanji character has multiple readings and each of these readings can be used alone or in conjunction with different kanji.

My question is, how is it that sometimes I see words that do not use any specified reading of the kanji, for example:

日 can be read as ひ, -び, -か,ニチ, ジツ

but in the word いっぴ (一日) it is used as ぴ


First of all, be aware that 一日 is usually read ichinichi, to count one day, or tsuitachi, when it refers to the first day of a month (as in "April the 1st"). But the phenomenon you spotted in ippi also happens to other words, so let's talk about that.

The simple explanation of ippi is that, when you join a ち (like いち) to ひ, the sequence ち-ひ usually becomes っぴ. So you see, the 日 is using its common ひ reading, after all. It's just that ぴ is considered to be a transformation of ひ (in this case, triggered by ち). More details follow.

The change from ひ to ぴ is an example of a sound change or euphony (Japanese: 音便{おんびん}). There are two important types of sound change which may happen when linking words into a compound:

Sequential voicing (連濁{れんだく}): The first kana character of the second word may get a dakuten ゛ mark. For example, 棚 たな → 本棚 ほんな. (In linguistics terms, the second word becomes voiced, except if it begins with an /h/, which becomes /b/. That's because /h/ was /p/ in Old Japanese, and a voiced /p/ is a /b/.)

If the kana cannot normally be written with a dakuten ゛, then no sound change happens. Even if it can, sometimes it doesn't change. It's a bit hard to predict when will rendaku occur (see here). Just be aware that it does happen, and with time you'll get a feel for it.

Gemination (促音便{そくおんびん}): When two words are linked, the first consonant of the second word may become doubled (gemini is Latin for "twins"). The doubled or "long" consonants are easier to see in rōmaji:

  • hatsu + 達 tatsuhattatsu
  • ichi + 個 koikko
  • ma + 白 shiro = masshiro

Gemination is often triggered by the first word ending with a つ or ち. In this case, the trailing つ or ち will become a "small tsu" (that is, the syllable will disappear, being subsumed into the new geminated consonant):

  • 発 はつ + 達 たつ → 発達 はたつ
  • 一 いち + 個 こ → 一個 い

A sequence of /k/ sounds also gets geminated, with one /k/-kana becoming a small tsu:

  • 悪 あく + 化 か → 悪化 あっか 

Sometimes no character becomes a small tsu; the small tsu seemingly pops out of nowhere:

  • 真 ま + 白 しろ → 真っ白 まっしろ

Finally, /h/-sounds (は,ひ,ふ,へ,ほ) are special. Whenever a /h/-sound gets geminated, it becomes a /pp/ sound; that is, the はひふへほ kana get a handakuten ゜ mark, becoming ぱぴぷぺぽ. (Again, this is because /h/ used to be /p/ in Old Japanese). This is what happened to 一日 as いっぴ, and it happens often for counters starting with 一 いち:

  • 一発 いっぱつ
  • 一匹 いっぴき
  • 一夫 いっぷ
  • 一片 いっぺん
  • 一本 いっぽん
  • 2
    Although this is implied, I think it would be helpful to state specifically in response to the question that the reason why, for example, ぴ is not a "specified reading" of 日 is because of this phenomenon. The ぴ in this case is actually a ひ, etc. – Ciaran Aug 26 '16 at 11:40
  • 1
    Nice answer. Do you (or anyone) know how to determine whether 一 as いっ is more likely to be イツ or イチ? (Here いっぴ could be イツ+ひ or イチ+ひ, right?) – Earthliŋ Aug 26 '16 at 15:32
  • 1
    I was talking about イツ as the kan on reading of 一 (as in 唯一 【ゆいいつ】), not about いつ as in 何時. – Earthliŋ Aug 26 '16 at 17:48
  • 1
    Oh, I see! Well, kan'on comes from Chang'an Late Middle Chinese, and go'on from Late Old Chinese; but the thing is, 一 “one” was *ʔit- in both, and entered Japanese as it- – the form it had in the Jesuit materials, ca. 16 century (cf. Frellesvig, section 11.4). I think the ichi/itsu pair doesn't come from real, distinct kan/go'on etymons, but this distinction is a late back-formation from it- – what Miller once called the "linguistic ghosts" of Japanese lexicography. Therefore, ippi actually comes from neither /itu/ nor /iti/, but from their ancestor and Old/Middle J form /it-/. – melboiko Aug 26 '16 at 18:27
  • 2
    As for 真っ白, some claim that it derives from prefix "ma" with reduplicated adjectives, which is shared with Filipino or Melanesian languages. This feature is well reserved in "ma aka aka" → まっかっか. – user4092 Aug 27 '16 at 0:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.