Furigana in Aozora Bunko: Why Some Books Have So Much More

Kusuyama Masao’s Momotarō on Aozora Bunko begins like this:

まいにち、おじいさんは山へしば刈《か》りに、おばあさんは川へ洗濯《せんたく》に行きました。

“Every day, the old man went to the mountain to cut brushwood, and the old woman went to the river to do the washing.”

And this is the opening of Natsume Sōseki’s Kokoro:

私《わたくし》はその人を常に先生と呼んでいた。だからここでもただ先生と書くだけで本名は打ち明けない。これは世間を憚《はば》かる遠慮というよりも、その方が私にとって自然だからである。

“I always called him Sensei. So here, too, I will only write Sensei and will not reveal his real name. This is less out of reserve towards the world than because it comes more naturally to me.”

Both come from Aozora Bunko’s text files, where the reading sits inside 《》. Momotarō gives a reading even for easy kanji like 刈 and 洗濯. Kokoro gives none for 常, 先生, 本名 or 世間, but it does for the rarer 憚.

If you read Aozora Bunko for long, you notice this: some books carry furigana on almost every word, others on very few. Here is how Aozora Bunko writes these readings (called ruby, ルビ), and why the amount changes from book to book, using Aozora Bunko’s own rules and the texts themselves.

Ruby is written with 《》

Aozora Bunko’s annotation list, in its section on ruby and characters attached like ruby, says ruby is written with 《》. The example is 青空文庫《あおぞらぶんこ》. In the XHTML version you read in a browser, this shows up as furigana above the word.

The plain-text files sometimes add ｜ to show where the ruby starts. According to the annotation list, when the characters under the ruby are of different kinds, such as kanji and kana, a ｜ goes in front of them. When they are all one kind, all kanji or all katakana, no ｜ is needed. Hiragana and katakana count as different kinds.

Aozora Bunko’s input manual, in section 4-1 on ruby, gives the example 武州｜青梅《おうめ》の宿, “the inn at 青梅”. The reading covers only 青梅, and the 武州 before it is also kanji, so ｜ marks where the reading begins. When the kanji under the ruby come after something that is not kanji, such as kana or punctuation, no ｜ is needed. Momotarō uses ｜ in the same way:

きじが一｜羽《わ》とんで来《き》ました。

“A pheasant came flying.”

Here the ruby わ belongs to 羽 alone, the counter for birds, and not to 一羽 as a whole.

Ruby is not always a reading. The annotation list also gives 霧の｜ロンドン警視庁《スコットランドヤード》: the kanji say “London Metropolitan Police”, and the ruby gives the foreign name it stands for, Scotland Yard.

Some things look like ruby but are not entered as ruby. When the printed book puts a small ママ beside a character, the manual explains that this is a note meaning “this looks like a misprint in the source, but it is left as it is”. It is entered as an inputter’s note, not with 《》.

The furigana come from the source edition

Every Aozora Bunko work is typed in from a printed book. In the manual’s opening essay, “本という財産とどう向き合うか” (How to treat the heritage of books), that book is called the 底本 (teihon), the source edition.

The input manual says Aozora Bunko aims at “an electronic file based on the source edition” and works on the principle of “no editing at will”. Ruby, too, is entered as the source edition has it. When an inputter corrects a reading that looks like a misprint, the proofreading manual has them leave a note of what the source edition said, as in 籠《ざる》［＃ルビの「ざる」は底本では「さる」］: the ruby ざる was さる in the source edition.

So the amount of furigana depends less on the work than on the source edition. Momotarō’s source edition is a collection of myths and folktales, 日本の神話と十大昔話; Kokoro’s is a paperback edition of the novel. Both are named at the end of each file, after 底本：. A book with many readings was a printed book with many readings.

The same story, two versions

Aozora Bunko sometimes holds one work in more than one spelling. The manual explains that works written in old kanji and old kana are registered in their original form (旧字旧仮名), in modern spelling (新字新仮名), and in a middle form with modern kanji but old kana (新字旧仮名). Different versions can have different source editions.

Miyazawa Kenji’s The Restaurant of Many Orders has a modern-kana version and an old-kana version. The old-kana version is based on 宮沢賢治全集8, a volume of his collected works. Here is the same moment in both, when the two hunters, lost in the mountains, talk about turning back.

The modern-kana version:

「ぼくはもう戻《もど》ろうとおもう。」

「そいじゃ、これで切りあげよう。なあに戻りに、昨日《きのう》の宿屋で、山鳥を拾円《じゅうえん》も買って帰ればいい。」

　風がどうと吹《ふ》いてきて、草はざわざわ、木の葉はかさかさ、木はごとんごとんと鳴りました。

The old-kana version:

「ぼくはもう戻らうとおもふ。」

「そいぢや、これで切りあげやう。なあに戻りに、昨日の宿屋で、山鳥を拾円も買つて帰ればいゝ。」

　風がどうと吹いてきて、草はざわざわ、木の葉はかさかさ、木はごとんごとんと鳴りました。

“I think I’ll go back now.” “Then let’s call it a day. On the way back we can buy ten yen’s worth of mountain birds at yesterday’s inn.” “The wind came gusting, the grass rustled, the leaves rattled, and the trees creaked.”

The readings for 戻, 昨日, 拾円 and 吹 appear only in the modern-kana version. Nearby, 腹は空《す》いてきた (“I’m getting hungry”) and 兎《うさぎ》 (“rabbit”) have readings in both. Even within the modern-kana version, 戻 gets its reading only in the first 戻《もど》ろう, and not in the なあに戻りに that follows. Which Japanese Classics Can You Read at Each JLPT Level? counts the furigana in both versions.

Furigana in versions rewritten into modern spelling

When a work exists only in old kanji and old kana, an inputter may rewrite it into modern spelling. Aozora Bunko’s guidelines for rewriting works into modern spelling have several items about furigana:

• Readings are added, as needed, to the author’s own unusual kanji spellings, and to pronouns, adverbs, conjunctions and auxiliary verbs that are rarely written in kanji today.
• Readings are added, as needed, to foreign proper names and loanwords written in kanji.
• Readings are added to words that are hard to read or easy to misread.
• From a source edition with readings on every kanji (総ルビ), some readings may be left out.

In these versions, the furigana are no longer the same as in the source edition. The guidelines require a note at the end of the file when this rewriting was done, so the end of a file tells you which book it came from and how it was entered.

What Is Aozora Bunko? covers the library itself, Reading Old Kana Spelling covers old kana, and When to Stop Using Furigana covers letting furigana go.

Reading ruby in Niho

I make Niho, a free reading app for iPhone and iPad with more than 17,000 Aozora Bunko works and a built-in dictionary. It has no ads and no in-app purchases.

Niho shows Aozora Bunko’s 《》 ruby as furigana above the words. Furigana appear only where the source edition has them. You can turn them on and off with Show furigana in Reading settings, under the Aa button on the reading screen.

A book’s detail page shows its spelling, such as Modern kana or Old kana. The shelves in Discover list the modern-kana versions, and old-kana versions can be found by search. For a kanji with no reading, long-press it and tap Look up to see its reading and meaning. Getting Started with Niho walks through your first book.