The Writing Revolution

by Amalia E Gnanadesikan

  A second stumbling block was honorifics. Speakers of Japanese showed respect to people they considered their social superiors by using a special set of pronouns and inflectional suffixes when addressing them. These distinctions could not be written in Chinese, but they were important to the Japanese. To solve this problem, certain characters were given peculiarly Japanese interpretations as honorifics. A style of “modified Chinese” developed, with Japanese honorifics and a more Japanese word order. The Chinese verb placement was confusing to Japanese readers and writers, and in informal works such as diaries and personal letters they tended to slip into a more natural verb-final style. However, it was not until the time of the Kamakura shogunate (1192–1333), during the feudal period, that official documents were allowed to stray from correct, Classical Chinese and began to be written in modified Chinese.

  Works written in Chinese could be read back either in Chinese or, with some mental gymnastics, in Japanese. The individual characters could be pronounced either in Chinese or as Japanese words of equivalent meaning. For example, , “mountain,” which is shn in modern Mandarin, could be pronounced in Japan either in Chinese or as the native Japanese word yama. A full sentence of characters, if read back in Japanese, had to be reorganized to get the correct word order, and the reader had to infer the appropriate Japanese inflections and particles. Because of this flexibility of reading, it can be difficult to determine precisely what language early Japanese documents were intended to be read in. The modified Chinese style, with its Japanese honorifics and un-Chinese word order, seems to have been intended to be read in Japanese, or perhaps a Japanese–Chinese hybrid. Poetry that rhymed was clearly meant to be Chinese, as Japanese poetry did not rhyme. Other documents could have been read either way. However, many words relating to philosophy, education, and high culture were borrowed into Japanese from Chinese, so there was only one way to say the word, no matter which language one was supposedly reading in.

  By the ninth century it is clear that many Chinese texts were being read in Japanese, as priests studying Buddhist texts began to annotate the characters with dots to show which words received which particles, helping readers to construct Japanese sentences out of Chinese text. This was no doubt cumbersome, but easier to do with logographic characters than with a phonological script. Many Japanese learned to read Chinese texts without actually learning Chinese.

  From these texts, words flooded into educated Japanese from Chinese in much the same way that Latin and Greek words entered English and many other European languages. Yet Chinese and Japanese are even more different from each other than English and Greek, and the newly borrowed words had to be altered to fit Japanese. Chinese morphemes were generally each a single syllable, but the Japanese language allowed only very simple syllables. So some Chinese syllables were simplified and others were spread out into two syllables. Chinese words were stripped of their distinctive tones, while the phonemes they contained were converted into the nearest Japanese equivalents. These AD aptations caused rampant homophony in Japanese. A particularly egregious case is ka, which is the Chinese-based (Sino-Japanese) pronunciation of 31 different characters on the modern list of 1,945 commonly used kanji, as Chinese characters are known in Japan. Modern Mandarin pronunciations of the same characters are (in pnyn) xià, xiá, ji, jià, ji, hu, huà, huá, gu, guò, gu, hu, k, k, hé, g, and gè.

  The borrowing of Chinese words occurred in three major waves. The first began with the introduction of Buddhism, traditionally dated to AD 552 and credited to a missionary from Korea. The Japanese pronunciation of the new religious vocabulary was probably adapted from that of southern Wu Chinese dialects of the time (see appendix, figure A.4). The second wave took place during the Nara period (ad 710–93), a time of stable imperial government built around Confucian ideals and state-sponsored Buddhism. Students were sent to China to study, and government officials were sent on diplomatic missions. They brought home many new words associated with Confucian philosophy, government affairs, and secular education. The Japanese pronunciation of this large set of Chinese loanwords was adapted from the standard dialect of the ruling dynasty of the time, the Tang.

  The third and smallest wave of borrowings occurred in the fourteenth century with the arrival of a new sect of Buddhism, Zen. These words are mostly concerned with Zen, and their pronunciation in Japanese was probably AD apted from their fourteenth-century pronunciation in Hangzhou, in southern China.

  The effect of these multiple borrowings at different times and from different kinds of Chinese was to permanently complicate written Japanese. When kanji are read in Japanese each character may have up to three Sino-Japanese pronunciations, known as on readings, as the morpheme may be part of loanwords from up to three phases of borrowing, taken from three different Chinese languages. Additionally there is the kun reading – the native Japanese pronunciation of a word of the same meaning. Luckily, the majority of kanji have only one Sino-Japanese on pronunciation. In China characters also have many different pronunciations, but the variation there is from one dialect to another – within a dialect, a character usually has a single pronunciation. In Japan, the particular pronunciation to be used depends on context.

  Sino-Japanese words, including new compound words made in Japan out of Sino-Japanese parts (many of which have since been borrowed back into China), now comprise about half the words in Japanese. Most simple, commonly used words remain Japanese, however. Compared to native words, Sino-Japanese words are considered more formal, technical, and precise, just as words of Latin and Greek origin are in English. In English, a word like water is an everyday, native English word. When the concept of “water” is used to make up scientific compound words, however, we use the Greek hydro- (in words like hydrology and hydroponics), or Latin aqua- (in words like aquatic and aquarium). If we were writing English like Japanese, the free-standing word water as well as hydro- and aqua- would be all written with the same character. We would then know that a word of a single character was to be read back as a native word (water), while the ones in compound words of technical vocabulary would receive a foreign pronunciation (hydro- or aqua-) depending on the topic of the word and the morphemes it was compounded with (-ology versus -ium).

  The Chinese language left a permanent mark on Japanese, not only in its writing system and its vocabulary, but also in its phonology. Chinese words were adapted to fit Japanese syllables, but they also exerted their own pressure on Japanese syllables, creating what are known as heavy syllables, which contain either a long vowel or a final consonant: CVV (also noted CV) or CVC. The types of closed (CVC) syllables in Japanese are still extremely restricted, but it is due to Chinese influence that they exist at all.

  Meanwhile, as the Japanese intellectual class blossomed during the Nara period, writing in Chinese or even modified Chinese was found to be inadequate for certain purposes, even if texts could be translated into Japanese on the fly by an adequately nimble reader. Writers could express most of their ideas in Chinese, especially as they owed so much of their intellectual culture to the Chinese, but what about their poetry? Poetry, like names, is deeply rooted in sound, not just meaning. The form of poetry is untranslatable between languages as different as Chinese and Japanese. To write Japanese poetry, therefore, required a way of writing Japanese. The initial solution to this problem was to expand the rebus technique used to write Japanese names: Chinese characters were used for their syllabic values, not their meanings. A syllabic system of writing known as man’ygana emerged in the eighth century, named after the Man’ysh collection of poetry compiled around 759. Early written poetry, such as that found in the Man’ysh, used a combination of characters used for their kun (native Japanese) readings and a syllabic use of characters for their sounds alone, the man’ygana syllabary. Texts in this logosyllabic style faithfully represented the Japanese language, including all its particles and inflections.

  The man’ygana syllabary was large and inefficient. The syllabic value of a character could be
derived from either its on reading or its kun reading. A number of different characters could be used for the same syllabic sound. Twelve different characters could spell ka, for example. This is a lot, but fewer than it could have been, considering the 31 kanji with that pronunciation. In the Man’ysh, 480 characters were used for their syllabic on values, and a smaller number for their syllabic kun values, all for the roughly 90 different Japanese syllables of the time.

  The man’ygana of early Japanese poetry is remarkably like Akkadian cuneiform, despite the lack of any visual similarity or historical connection. Both used a bulky logosyllabary, with phonological values being supplied by two different, unrelated languages (Sumerian and Akkadian being as different as Chinese and Japanese). Like Akkadian cuneiform, the Japanese system was cumbersome but worked: Japanese had become a written language. Nevertheless, writing in Japanese did not enjoy high prestige. Man’ygana was reserved for poetry and proper names; other eighth-century texts were in either proper Chinese or modified Chinese. Within the restricted field of poetry, however, the Man’ysh had a tremendous literary influence, and its style, containing virtually no Chinese loanwords, became the model for native-style poetry. In the aesthetic context of poetry, native vocabulary was valued, in sharp contrast with the importance of Chinese in formal documents. Where people’s hearts were involved, they wrote in Japanese.

  The need to record the Japanese language accurately also arose in the cases of imperial rescripts and Shinto prayers, which were written in a style known as senmygaki (imperial rescripts are senmy in Japanese). The Shinto prayers were of native Japanese origin and were supposed to be repeated accurately, word for word. Similarly, the emperor’s words were taken down so as to be read back exactly, and of course the emperor spoke Japanese. In senmygaki, kanji were used with their kun readings, with the particles and inflection written in smaller man’ygana characters. The result was the first truly Japanese prose.

  The Nara period was succeeded by the Heian period, which lasted from 794 until 1192. During the first hundred years, Japanese literature languished, but the tools with which to create it were refined, and the two Japanese syllabaries still in modern use were born.

  Japan’s first native syllabary, hiragana, developed out of a cursive version of man’ygana characters. (The shared element -gana of hiragana and man’ygana is derived from kana and means “syllabary.” In the formation of certain compound words the k becomes a g – a “hard” [g] sound – while in others it remains k.) The cursive man’ygana characters were reduced to simple, rounded shapes of only a few strokes apiece. The result was hiragana, “smooth kana.” Even an inexperienced eye can pick out the hiragana in a page of modern Japanese, as the signs are noticeably more curved – and usually simpler – than the accompanying kanji. The word kanji, for example, is written (in kanji) , while kana (in hiragana) is the much simpler . The circular element at the bottom of the na sign would never occur in (noncursive) kanji, where only slight curves are permitted.

  According to tradition, hiragana was invented by the sainted Kb Daishi, founder of the Shingon sect of Buddhism. Probably the tradition is not entirely accurate, as hiragana seems not to have fully taken shape within Kb Daishi’s lifetime, from 774 to 835. Little documentation of the development of hiragana remains, but it seems to have been an evolutionary process rather than the invention of an individual or of a moment. Different ways of simplifying man’ygana characters were tried, and at first some of the man’ygana syllabary’s redundancy was inherited, with more than one sign for the same syllable. Eventually, a lean syllabary of 50 signs was achieved.

  At nearly the same time, another form of kana was being created out of man’ygana characters. Buddhist monks and their students, poring over Chinese texts, would annotate the texts, showing the pronunciation of unfamiliar kanji, or recording the particles and inflections needed to turn the sentences into Japanese. The spaces between characters were not large, however, and students were often under time pressure as they took notes during lectures. In response to these pressures, they began writing only part of the man’ygana characters as a kind of abbreviation, giving birth to “partial kana” writing, or katakana. Abbreviated writing had been practiced before, both in Japan and in China, but the usefulness of this new kana, not just as shorthand but as a way of writing Japanese rather than Chinese, could not be overlooked. Like hiragana, katakana took a while to become standardized, with rival forms – varying in which part of the character got simplified, or which character got simplified – existing for some time.

  Because katakana was abbreviated from standard (noncursive) characters, it retains the angular shape of noncursive kanji, but is much simplified. is “hiragana” in hiragana, while the same word appears more angular in katakana as . In some cases the hiragana and katakana signs were based on the same man’ygana character. Sometimes this is obvious: hiragana and katakana (both ka) are derived from the character, meaning “to add,” which was also pronounced ka in its on reading and was used as one of the many man’ygana characters for that syllable. At other times the derivation from the same character is not obvious: hiragana and katakana (me) are both derived from , pronounced me in its kun reading, meaning “female.” In other cases, due to the redundancies available in the man’ygana syllabary, equivalent characters in hiragana and katakana were taken from different characters: hiragana ha (also used for the particle wa) is from ha, meaning “wave,” while the corresponding katakana sign, , is taken from , meaning “eight” and pronounced hachi. Figure 7.1 shows the hiragana and katakana syllabaries.

  Unlike Akkadian or Mycenaean Greek, Japanese is well suited to be written with a syllabary. At the time the kana were being developed, Japanese consisted of only light syllables: CV (or just V). In their stabilized and standardized forms, hiragana and katakana each comprised 50 signs. Of these, 45 are still used. A forty-sixth has since been added, and new uses of some signs have been developed, to account for changes in the Japanese language since the ninth century. Due to the influence of Chinese, modern Japanese syllables may contain a long vowel or could be closed in one of two ways. First, a syllable may be closed by a nasal sound. If the nasal is followed by a consonant, it will be pronounced in the same part of the mouth as that following consonant, as in familiar Japanese words like tempura (itself a loanword from Portuguese) and Honda. If the nasal is not followed by a consonant it is pronounced as a rather indeterminate nasal sound far back in the mouth. The other way of closing a syllable in Japanese is by doubling the consonant that begins the next syllable. In a word like Hokkaido, for example, the first k closes the first syllable. The mouth remains closed, prolonging the k, until it is time to pronounce the vowel of the second syllable.

  The forty-sixth sign – hiragana and katakana – represents the adopted syllable-final nasal. A small version of the tsu syllable is used for the other kind of closed syllable: in hiragana, or in katakana, indicates that the upcoming consonant is doubled. Another special application of kana symbols occurs for what the Japanese call “twisted” sounds and linguists call palatalized. Palatalized consonants sound almost as if they are followed by a y ([j] in the International Phonetic Alphabet), and so are transliterated kya, kyu, kyo, etc. In kana they are written with two symbols, the second one written smaller to emphasize that there is actually only one syllable. So the syllable kya is written as though it were ki-ya: . The representation of palatalized sounds did not begin until the middle of the Heian period, but it is not clear whether these syllables developed under the influence of Chinese, or were native but left out of the original kana in the interests of achieving a smaller, more easily memorized set of signs.

  Figure 7.1 The Japanese syllabaries, with hiragana on the left and katakana on the right. The Romanization follows the Hepburn style, with IPA interpretation where needed. The basic syllabaries are above the double line, secondary symbols with diacritics below. A smaller version of the tsu character is used for the first part of a double consonant.

&nb
sp; The 46 basic signs do not distinguish between syllables that begin with g versus k, s versus z, or t versus d. These pairs of sounds are distinguished by the phonological property of voicing: in [g] the vocal cords are vibrating, creating a person’s voice, while in [k] the vocal cords do not vibrate, and the sound is effectively whispered. So, for example, ka and ga were originally both in hiragana. Evidence from man’ygana shows that Japanese of the Heian period did distinguish between voiced and voiceless sounds, but it is not unusual for a syllabary to ignore this difference in return for a smaller syllabary. Linear B did the same, and even English fails to mark one voicing distinction, spelling as th both the voiced [] of either, and the voiceless [θ] of ether.

  In order to distinguish syllables with the voiced sounds g, z, d, and b from their voiceless counterparts, diacritical marks were added during the feudal period. So is ka in hiragana and is ga. Also added was a diacritical mark to distinguish syllables beginning with p from those beginning with h. At some point in the history of Japanese, [p] came to be pronounced as [h] (except before [u], where it is pronounced [Φ], a sound similar to [f]). The result was that [b] came to be considered the voiced equivalent of [h]: is hiragana ha, and is ba. A special diacritic is nowadays used to make pa: . The [p] sound is now limited to doubled consonants (it is the doubled version of h), consonants occurring after the nasal, onomatopoeic words, and foreign loanwords.

  Hiragana and katakana both evolved in the ninth century from the same source, man’ygana. However, they were developed, and continued to be used, in different environments. Katakana arose in the austere, masculine environment of Buddhist scholarship; its use in the Heian period was restricted to men. By the end of the Heian period, katakana had replaced the small man’ygana characters for Japanese particles and inflections in the senmygaki style of writing, yielding a form of written Japanese that mixed kanji and katakana.