Japan: Chinese characters and kana

A script from the mainland

Japan encountered Chinese writing before it began to use it. According to Chinese annals, in AD 57 Emperor Guangwu of Han granted a gold seal to an envoy from the Japanese land of Na. Objects bearing Chinese characters appear on the islands from the 1st century, but by the prevailing view regular writing began only in the 5th century. Both the Kojiki and the Nihon Shoki state that writing and Chinese books were brought by scholars from the Korean kingdom of Baekje; there is no direct evidence, but scholars generally accept it.

The oldest inscriptions made in Japan are Chinese texts, probably written by immigrant scribes. The best known is the sword from the Inariyama burial mound, excavated in 1968: X-rays in 1978 revealed a gold-inlaid inscription of at least 115 characters. It is dated to a shinkai year, by the prevailing interpretation 471, and the Japanese personal names in it are written phonetically, with characters used only for their sound. Similar inscriptions appear on the Eta Funayama sword and the Suda Hachiman Shrine mirror.

The decisive impulse came with Buddhism, which reached Japan from Korea in the 6th century. It brought sacred texts (sutras) written in Chinese, temples became centres of learning, and Chinese became the language of government, religion and scholarship, much as Latin was in medieval Europe. Texts written in Chinese are called kanbun in Japan, and a large corpus of them survives.

Reading Chinese as Japanese

Japanese differs from Chinese in structure. A Chinese word is usually a single uninflected syllable and the sentence runs subject – verb – object. Japanese words often have several syllables, verbs are conjugated with suffixes, grammatical relations are marked by particles placed after words, and the verb stands at the end of the sentence. The Japanese therefore learned to read Chinese text directly as Japanese, a method called kanbun kundoku: the reader rearranged the words into Japanese order, added Japanese endings and particles, and read the characters as Japanese words of the same meaning.

Small marks added to the text (kunten) served this purpose: dots at the edges of characters, return marks (kaeriten) showing the order in which the characters are to be read, and later small syllabic signs for the endings. The earliest securely dated Japanese examples of return marks are from 783 and 788; the older of the two may reproduce glosses from the Korean kingdom of Silla. The method has survived to this day: Japanese secondary schools teach kanbun as part of language education, and a single standard for the marks has applied since 1912.

For the same reason most characters have two kinds of reading in Japanese. The on reading (on'yomi) imitates the Chinese pronunciation of the character at the time the character and its word were borrowed, and is used mainly in words of Chinese origin. The kun reading (kun'yomi) is the native Japanese word matching the character's meaning. The character 人 “person” is read hito in Japanese but jin or nin in Chinese compounds; the character 行 is read i in iku “to go”, okona in okonau “to carry out”, gyō in gyōretsu “line”, kō in ginkō “bank” and an in andon “lantern”. Because Chinese words were borrowed at different times and from different regions, a character may have several on readings.

Characters for sounds: man'yōgana

Names, poems and Japanese endings could not be written in Chinese. Scribes therefore began to use Chinese characters for their sound alone, regardless of meaning, as the Chinese and Koreans had earlier done with foreign place and personal names. In Japan this grew into a complete syllabic script called man'yōgana. Characters were read either by their Chinese pronunciation (for example 波 for ha) or by a Japanese word (the character 女 “woman” for the syllable me). There were no fixed rules: by the end of the 8th century some 970 different characters were used for about 90 syllables of Japanese, and the syllable ka alone could be written with 加, 可, 何, 架, 香 or 迦, among others.

The oldest Japanese books were produced at the imperial court in Nara in the early 8th century. The Kojiki (“Records of Ancient Matters”), completed in 712, tells of the creation of the world, the gods and the first emperors; its prose is written in Chinese characters with a Japanese-influenced structure, while its 112 songs are written phonetically. The Nihon Shoki (“Chronicles of Japan”) of 720 is written in Chinese on the model of Chinese histories, but its 128 songs are also in man'yōgana.

The script takes its name from the Man'yōshū (“Collection of Ten Thousand Leaves”), the oldest surviving anthology of Japanese poetry. It has 20 books and over 4,500 poems, among them 4,207 short poems (tanka) and 265 long ones (chōka); the last datable poem is from 759. Its last compiler was, by the prevailing view, the poet Ōtomo no Yakamochi. The poems were written by emperors and courtiers but also by frontier guards and people from the eastern provinces, so the collection records eastern dialects too. The Man'yōshū mixes characters used for sound with characters used for meaning, and it is still difficult to read.

Writing outside the court is documented by mokkan, wooden tags and slips. Officials wrote notes, letters and lists on them and attached them to shipments of taxes. More than 150,000 from the late 7th and early 8th centuries have been excavated. Their language is more colloquial than that of the poems, and they show that a fairly large number of officials could already write. Three of the tags carry poems that also appear in the Man'yōshū.

Old Japanese

The language of the Nara period (710–794) is called Old Japanese. Like modern Japanese it was agglutinative: suffixes and particles were added to the stem, the verb came at the end of the sentence and modifiers preceded the noun. Syllables were simple, usually a consonant and a vowel, and a word could not begin with r or a voiced consonant. Modern h was originally pronounced p (the syllable ha, written with 波, sounded pa); whether it was still a p in the Nara period or already a bilabial f is disputed.

Careful analysis of man'yōgana shows that the scribes of the oldest texts distinguished 88 syllables, more than later Japanese. For the syllables ki, hi, mi, ke, he, me, ko, so, to, no, yo, ro (and in the Kojiki also mo) they always used one set of characters for certain words and another set for other words. The phenomenon is called jōdai tokushu kanazukai, and the two series are labelled kō (甲, A) and otsu (乙, B), the first two of the ten Chinese “heavenly stems”, i.e. roughly “first” and “second”. The linguist Shinkichi Hashimoto described it systematically in 1917; the distinction disappeared during the Heian period.

What exactly the difference meant is debated and there is no consensus. The traditional interpretation, put forward by Kyōsuke Kindaichi in 1938, is that Old Japanese had eight vowels instead of today's five: besides i, e and o there were three more, pronounced further towards the centre of the mouth. Others, especially since the work of Roland Lange in 1968, explain the difference as diphthongs or as the glides y and w before or after the vowel. Internal reconstruction points, on the contrary, to an earlier stage with fewer vowels.

Most texts record the language of the court at Nara, Western Old Japanese. Poems from the eastern provinces in Books 14 and 20 of the Man'yōshū show an eastern dialect with different forms; according to some scholars (Alexander Vovin and others) it also contains loanwords from Ainu. Some of its features survive in the Hachijō language, spoken on the islands of Hachijōjima and Aogashima south of Tokyo.

Kana: a script from cursive and from fragments

Writing Japanese with dozens of complex characters was slow, and during the 9th century the characters used for sound were simplified. Two syllabaries arose, together called kana. Hiragana (“plain kana”) grew out of man'yōgana written in a very fluid cursive (sōsho): each hiragana sign is a simplified whole Chinese character, for example あ from 安. Katakana (“partial kana”) arose by keeping only part of a character written in regular script, usually its first or last few strokes: カ is the left half of 加.

Katakana was created by Buddhist monks in Nara. They wrote it small between the lines of sutras written in Chinese to indicate the Japanese reading and endings, so katakana began as a reading aid. Koreans used similar abbreviated parts of characters (the gugyeol script), and some katakana signs resemble them in shape. The two kana sometimes derive from the same character (か and カ from 加), sometimes from different ones (る from 留, ル from 流).

Tradition credits the invention of kana to the Buddhist monk Kūkai (774–835), founder of the Shingon school, who returned from China in 806 with knowledge of the Indian Siddhaṃ script. This cannot be proved; kana developed gradually and by many hands. Indian phonetics did, by the prevailing view, shape the order of syllables in the gojūon (“fifty sounds”) table: the vowels a, i, u, e, o and the consonants k, s, t, n, h, m, y, r, w follow one another much as in the Sanskrit alphabet. The oldest tables of this kind date from the 11th century; the present order became fixed only later.

For a long time each syllable could be written with several signs. Hiragana received a single standard form only in the school reform of 1900, which chose one variant for each syllable; the others are now called hentaigana (“variant kana”) and appear mainly in calligraphy and on shop signs.

The writing of court ladies

In the Heian period (794–1185), named after the new capital Heian-kyō, today's Kyoto, men's official writing remained Chinese and the regular script of characters, called otokode (“men's hand”). Hiragana instead took hold in private letters and poetry, and because it was used above all by women, who were usually denied a Chinese education, it was called onnade, “women's hand”. It allowed court ladies to write as people spoke, and they created the masterpieces of Japanese prose.

Yet hiragana was not only for women. In 905 Emperor Daigo ordered the compilation of the Kokin Wakashū (“Collection of Ancient and Modern Poems”), the first anthology of Japanese poems (waka) compiled by imperial command. One of its compilers, the poet Ki no Tsurayuki, wrote the Japanese preface in kana (kanajo); the Chinese preface (manajo) is by Ki no Yoshimochi. Around 935 the same Tsurayuki wrote the Tosa Diary, describing the journey back to the capital from Tosa Province, where he had been governor. He wrote it in hiragana, as if it were written by a woman in his party.

The summit of Heian prose is the Tale of Genji (Genji monogatari) by a lady-in-waiting known to us only by the nickname Murasaki Shikibu. The 54-chapter novel, written in the early 11th century mostly in hiragana, tells of the life and loves of the emperor's son Genji and his descendants, and is often called one of the first novels in world literature. From about the same time comes the Pillow Book (Makura no sōshi) of the court lady Sei Shōnagon. Both authors knew Chinese writing, but they did not have to write in it.

The iroha poem

For a long time the order of syllables was learned not from a table but from the iroha poem, named after its first three syllables. The poem is a pangram: each of the 47 syllables of the time occurs in it exactly once, including wi (ゐ) and we (ゑ), which disappeared from standard spelling only in 1946. The syllable n (ん) is missing, because it did not yet have a sign of its own and was written like mu.

Tradition attributes the poem to Kūkai, but the attribution is no longer accepted. The poem no longer distinguishes the vowel e from the syllable ye, which earlier Japanese still had, and it is therefore most often dated to the 10th or 11th century. The oldest surviving copy is from 1079. Its content is Buddhist: by the usual interpretation it paraphrases a verse on the impermanence of all things from the Nirvana Sutra.

Dictionaries ordered syllables by the iroha until modern times, and it is still sometimes used to number items in a list. In music the syllables i, ro, ha, ni, ho, he, to (in katakana イロハニホヘト) name the notes A to G, so ハ長調, ha chōchō, is C major. The usual order of syllables today, however, is gojūon.

Three scripts in one sentence

Modern Japanese text mixes three scripts. Kanji write the roots of words, hiragana endings, particles and words without a common character, and katakana loanwords, foreign names, onomatopoeia and the names of animals and plants. In the sentence 私はビールを飲みます (watashi wa bīru o nomimasu, “I drink beer”) the characters 私 “I” and 飲 “drink” are kanji, the particles は and を and the ending みます are in hiragana, and ビール “beer”, borrowed from Dutch, is in katakana. Czech names are written in katakana too: プラハ Puraha is Prague, ヴァーツラフ広場 Vātsurafu hiroba Wenceslas Square.

After 1945 the script underwent a major reform. In 1946 kana spelling was brought into line with actual pronunciation (gendai kanazukai) and the government limited the characters for general use to 1,850 (tōyō kanji); many characters were given simplified shapes (shinjitai). Since 1981 a recommended list, the jōyō kanji, has replaced the binding limit; it was extended to 2,136 characters in 2010, and pupils learn 1,026 of them by the end of the sixth grade. The largest character dictionaries list over 50,000, but tens of thousands of them occur only in old texts.

The Japanese first met the Latin alphabet (rōmaji) in the 16th century through Portuguese traders and missionaries; the first romanization of Japanese, based on Portuguese spelling, was devised around 1548 by a Japanese Christian named Anjirō. Today the most common system is Hepburn romanization, named after the American missionary James Curtis Hepburn and adapted to English (shi, chi, tsu, ji); alongside it there is the official Kunrei system (si, ti, tu, zi). Czech has its own transcription based on Czech pronunciation (ši, či, cu, dži, ja, jo), which is why Czech writes Murasaki Šikibu and Kodžiki where English has Murasaki Shikibu and Kojiki.

The Japonic languages

Japanese belongs to the Japonic family. Besides Japanese and its dialects, the family includes the Hachijō language and the Ryukyuan languages of the Ryukyu Islands between Kyushu and Taiwan, such as Okinawan, Miyako, Yaeyama and Yonaguni. The Ryukyuan languages preserve some features older than those of the Old Japanese of Nara. They are not mutually intelligible with Japanese, but in Japan they have traditionally been regarded as its dialects. Most of them are now severely endangered.

No relationship between the Japonic languages and any other family has been proved. The most studied hypotheses link them to Korean or to the so-called Altaic languages, or posit Austronesian influence. Modern Japanese and Korean have very similar sentence structure, but there are too few shared words with regular sound correspondences, so most scholars do not consider a common origin proven.

Also unrelated is Ainu, the language of the indigenous people of Hokkaido, Sakhalin and the Kurils, which together with extinct Sakhalin and Kuril Ainu forms a small family with no proven relatives; it has its own card in the atlas. According to some scholars (Alexander Vovin and others), Eastern Old Japanese borrowed a number of words from it. Ainu is now often written in katakana extended with small signs for syllable-final consonants.