From Pimiko to Fimiko?
A history of Japanese writing before 1600
The history of Japanese writing is a relatively short but eventful one. In the span of only a couple of centuries the Japanese people went from having no writing system at all to having one of the richest and most complex. Fortunately, because Japanese writing is so young we know a great deal about its origins. The 807 AD work Kogojuui (“Gleanings from the ancient language”) stated in its preface that writing was not known to the Japanese people until Chinese script was imported via Korea. A 13th century commentary on Nihon Shoki, however, claimed that writing had existed in Japan since time immemorial, and the so-called “language of the gods” was supplanted by Chinese script. Debate raged on for centuries over which assertion was the correct one. There was a certain appeal to believing that the gods themselves had granted writing to the Japanese people, and so this belief came about long after everyone who knew the truth firsthand had died of old age.
The earliest record of writing in Japan is a record of correspondence between the queen Pimiko (c. 170-248 AD)—Himiko in Modern Japanese—and the Cao Wei kingdom in China. This is interpreted either to mean that Himiko and at least some of her court officials were literate, or that Himiko dictated to Chinese scribes who took up residence in her court (the second option is more likely). Either way, it is certain that literacy was not widespread during this era. Japan’s first two major works, Kojiki and Nihonshoki, state that Korean teachers from the kingdom of Paekche brought Buddhism and writing to Japan.
Writing became extremely popular under Empress Suiko (r. 592-628). Inscriptions from this time period include the Gangouji tray, the Gangouji Jouroku Sakyamuni halo, the Houryuuji Yakushi Buddha halo, and the Tenjukoku mandala. These inscriptions, as with all others from this time, include only proper nouns (with one exception). Overall they contain 60 of the estimated 87 syllables of Old Japanese (Modern Japanese has only 46 “legitimate” syllables). The remaining 27 were infrequently used and so were statistically unlikely to be included on any inscription. Other extant writings from before the Nara Era include the Taihou census of 702, the four poems of Jogu Shoutoku hou-ou teisetsu, various inscriptions, and various edicts and liturgies. If a writing used Chinese syntax it was called 純漢文 (junkanbun, “pure Chinese writing”) but if it used Japanese syntax it was called 変体漢文 (hentaikanbun, “perverted Chinese writing”). Beginning with the Heian Era the use of junkanbun declined over time, and especially from the start of the Kamakura Era when Japan was governed by warriors who had little time for such elaborate writing. There were, however, some well-educated samurai who mastered and used hentaikanbun—government edicts generally were written in hentaikanbun from the Kamakura Era to the Meiji Era.
Hentaikanbun was defined by its lack of standardization and by the 7th century there was at least one extant inscription written entirely in the Japanese word order. Most of the writings mentioned above, aside from the inscriptions, used a style known as 宣命書き (senmyougaki, “edict-writing”). This style used small phonograms for particles and inflections in addition to the regular-sized characters which were used as semantograms that represented the nouns, verbs, etc. These phonograms may have been included because edicts were meant to be read aloud; the phonograms aided the speaker in reading the Chinese script with Japanese word order and grammar.
We can reconstruct Old Japanese pronunciation fairly accurately by looking at the Chinese characters that ancient scribes used when writing Old Japanese. Writers of Old Japanese always used Chinese characters (or kanji) for their sound value with no regard to meaning. The kanji 乃, for example, means “therefore” or “whereupon” in Chinese but Japanese writers used it for the sound no, which is also the possessive particle in Japanese. (As an illustration, the section header at the top of this paper uses a string of kanji for their sound values in the same manner as Old Japanese, and the Modern way of writing is in parentheses following.) By the reign of Suiko, however, it was no longer a guarantee that kanji were used only for their sound values. Senmyougaki, in fact, did not use any kanji for sound values except for the small kanji being used for particles. By the 9th century kanji usage became inconsistent and could vary from writer to writer and from convention to convention. Sometimes a poem was recorded in two different books using two different writing systems. This is very helpful for confirming the pronunciations of characters.
Some works used two or more writing systems. Kojiki, for example, used senmyougaki for its prose section but the old phonetic system for its poetry. The Man-youshuu, an anthology of ancient poems, was written primarily in 戯書 (gisho, “playful writing” or “rambling writing”) which used characters either for their meaning or their sound, and furthermore could use either the 音読み (on-yomi, “sound reading”) or 訓読み (kun-yomi, “meaning reading”) that had developed by that time. Basically, gisho recorded Japanese in the most convoluted way possible and has been nearly impossible to read by a layman since the 11th century. Aside from its variant readings, gisho also included puns and plays-on-words that required lateral thinking. Possibly the best example of this is to be found in the pronunciation of the number 十六 (juuroku, “sixteen”) as 獅子 (shishi), meaning “lion.” The number four is pronounced shi, and four times four equals sixteen. To put it another way, the word was written as “sixteen,” read as “four four,” and understood as “lion.” Clever.
Although every Japanese work was written exclusively with Chinese characters, there were very few Chinese words in the pre-Nara works overall. Most of the Chinese (and Sanskrit) words that were used were Buddhist terms that had no Japanese equivalent. As time went on there were steadily more writers from different walks of life—a few of them used many foreign terms but most of them used none. Genji Monogatari, the most famous and well-respected Japanese work of the Heian Era, used only a few Chinese words, and only the ones that the intended audience was assumed to understand already. The use of Japanese vocabulary was simply called 大和言葉 (Yamato kotoba, “Japanese words”). When there was no Japanese word for something, writers either used existing Japanese words in a different context or coined new ones. This practice flourished during the Heian Era but declined afterwards. Before and after the Heian Era, but not during, new words were sometimes given kanji based only on sound. This practice is called 当て字 (ateji). The word 面白い (omoshiroi), for example, means “interesting” but consists of the kanji for “face” and the kanji for “white.” The meanings of those kanji are irrelevant in the word 面白い; the sounds just happen to fit.
When using a kanji for its sound value, Old Japanese writers could select from any number of kanji. For example, the sound a could be written using安, 阿, 悪, 亜, or 愛. The hiragana letter あ (a) was derived from the kanji 安 and the katakana letter ア (also a) was derived from阿. Therefore we know that whenever these kanji appear in an ancient work and they are being used for their sound value, that sound is always a. This consistency enabled Japanese scholars of later generations to infer correct readings with about 90% accuracy. Unfortunately Japanese was strongly influenced by Chinese at several points throughout history so not all pronunciations are preserved. Although阿 and亜 are still pronounced a (by the onyomi) in Modern Japanese, 悪is now pronounced aku and愛is pronounced ai. These do not represent natural changes in the language over time. Another effect that these waves of influence had was to increase the number of readings of certain kanji. Although most kanji have two or three readings (one or two onyomi and one or two kunyomi), some can have over a dozen. The character 安, for example, only has one onyomi, an (formerly a), and one kunyomi, yasu. 悪, on the other hand, has two onyomi, aku and o, and a staggering seven kunyomi: waru, a, niku, aa, ita, izukuni, and izukunzo. Fortunately, although the sounds have changed the characters are visually pristine. All Chinese characters are referred to as 漢字 (kanji, “Han [Dynasty] character”) because Classical Chinese characters were finalized by the end of the Han Dynasty.
The first time that Japanese was influenced by Chinese was in the 6th century when writing was imported. This first group of readings is called 呉音 (Go-on, “[Kingdom of] Wu sound”). An example of a go-on reading is myou for 明. Because the Koreans pronounced Chinese characters with a Korean accent, the pronunciations that the Japanese learned from them in the 6th century were not authentic. For this reason these readings are described as “imitations of Korean imitations.” The second wave of aural influence came during the T’ang Dynasty of China in the 7th-9th centuries; these readings are called 漢音 (Kan-on, “Han [Chang’an] sound”) and were brought directly from Chang’an by Japanese scholars. An example of a kan-on is mei for 明 as opposed to myou; both readings are still in use for different words, such as 明瞭 (meiryou, “clarity”) or 明星 (myoujou, “morning star”). The third major wave of influence occurred during the Sung Dynasty in the 14th-16th centuries; these readings are called 宋音 (Sou-on, “Sung [Dynasty] sound”). Most of these readings were imported piecemeal and used for specific words rather than systematically applied to kanji across many words; the sou-on reading min was created and applied to the kanji 明 to mean “Ming [Dynasty].” As with most other sou-on readings, the reading min has to this day only been used for the word for which it was created, namely, “Ming.” If not for these three Chinese onyomi, 明 would likely just have the native Japanese kunyomi aka and aki.
Japanese today is written in a mixture of kanji, the odd Roman letter, and a syllabary known as仮名 (kana). The word 仮名 literally means “informal name” or “temporary name,” alternatively written as 仮字, meaning “informal character” or “temporary character.” Kana is divided into two versions which have identical sounds and syllables, but are written differently. They are 平仮名 (hiragana, literally “ordinary kana”) and 片仮名 (katakana, literally “fragmentary kana”), In Modern Japanese hiragana is used a majority of the time whereas katakana is used primarily to write foreign loan-words. This rule, however, was not instituted until 1946; before then, there were no rules.
Japanese writers composed poems in Old Japanese using Chinese script, but they also wrote Chinese and studied Chinese poetry. The popularity of Chinese study varied through the ages. It reached its height of popularity in the early 9th century, during which the use of the aforementioned “small phonograms” in senmyougaki declined in use. With the fall of the T’ang Dynasty of China in 907, interest in Chinese quickly fell and Japanese composition experienced a dramatic resurgence. Along with this resurgence came the adoption of kana.
Kana was developed in the 9th century, with the oldest known writing dating to 828. Katakana characters were originally used only for particles and were called ヲコト点 (wokototen). Similarly to the small kanji used for particles in senmyougaki, these were utilized to aid reading in Japanese word order. Katakana was used at first on a person-to-person basis and gradually spread in popularity. Wokototen were written in red or white ink whereas the ordinary characters were written in black. Many students also began to write in only katakana while listening to lectures in order to take notes more quickly.
The two kana systems came about in their own way. Hiragana was derived from cursive versions of kanji. As cursive writing was already common in China and Japan, these characters were called “ordinary kana” as mentioned before. Katakana, as the name suggests, was taken from fragments of kanji. The concept of writing only part of a character (e.g. 寸 for 村) had been in use in China for centuries, but was always a placeholder—never officially recognized and never standardized. It was only in Japan that the practice became universal. By the end of the Heian Era there were eight different schools of wokototen.
In Figure 1 each kana letter is shown on the left side of the cell and its corresponding kanji on the right (in the hiragana chart there is also a third “intermediate” character). If we look at a kanji that a hiragana and katakana letter have in common we can see the two different approaches to its adaptation. The kanji 久 (ku), for example, was written in cursive to derive the hiragana letter く and a part of the kanji was removed to make the katakana letter ク. There were originally well over 50 kana, but many of these slipped away; most of them almost immediately and others in the centuries to come. The famous pangram poem “Iroha,” written no later than 1079, shows us that by then the syllabary included the 46 letters of modern kana (except for ん which was represented by む) as well as wi and we–yi and ye had fallen into disuse. By the year 1000 there was often confusion between い (i), ゐ (wi), and ひ (hi); え (e), 𛀁 (ye), and へ (he); お (o), を (wo), and ほ (ho); わ (wa) and は (ha); and う (u) and ふ (fu). Writers in the Kamakura and Muromachi Eras attempted to follow established pronunciations but with little success because some sounds had been lost and others had changed. Fujiwara no Teika was the first person to have any success with following and recommending Heian Era standards for kana usage; his one mistake was in confusing お (o) and を (wo). But to be fair, even in modern-day Japan the letter を is more often than not pronounced o. The 14th century scholar Byoua studied Teika’s work and wrote the book Kanamojizukai (“How to use kana”). Because Teika had achieved an authoritative status by then, his spelling system became more or less dominant.
The creation of hiragana is traditionally attributed to the priest Koubou, but this is doubtful. It is more likely that its creation was a collaborative effort, and it is known that there were multiple competing versions of kana letters for centuries afterwards. Katakana had a certain uniformity by the 10th century, but hiragana, being cursive and associated with poetry and fiction, had far less rigidity. Aside from the appearance of kana letters, deciding which set to use was also up to the author. Different writers had different ways of writing, and as time went on these changed. Of the three earliest copies of the book Sanbou e-kotoba, one is in hentaikanbun, one is in hiragana, and one is in a mixture of katakana and kanji. For the first few hundred years, hiragana was considered feminine and was called 女手 (onnade, “female hand”) whereas katakana was called 男手 (otokode, “male hand”). Women writers used only hiragana and male writers wrote in either katakana, kanji, or a mixture—the mixture was called 仮名交じり文 (kanamajiribun, “mixed kana writing”). The male author Ki no Tsurayuki had to use a pen name in order to write and publish works in hiragana. When men corresponded with women, however, they wrote only in hiragana. Some conventions were universal, however. During the Heian Era letter-writing came to use the honorific term 候 (sourou, “to attend”), and the style that used this honorific was known as 候文 (souroubun). Souroubun continued to be used as the official letter-writing style up until the mid-20th century. The narrative prose of Heian Era writings used honorifics, as opposed to modern-day narratives that virtually never do.
Just as with kanji pronunciations, kana has also experienced subtle changes over the years. Most scholars believe that the “H” row was originally pronounced “P,” i.e. pa, pi, pu, pe, po; but by the Muramachi Era the row had become “F,” i.e. fa, fi, fu, fe, fo. That is why the title of this paper is “From Pimiko to Fimiko.” It is a great blessing that we have external proof of these changes. Portuguese Jesuit missionaries quite actively published a wide array of books throughout the 1590s, including works published in Japanese but using the Roman alphabet. They therefore captured an accurate snapshot of Japanese pronunciations at the end of the 16th century. That is genuinely exciting.
Kana originally had no 濁点 (dakuten, “marks for voiced syllables”). The marks ゛and ゜were adopted by the end of the Heian Era but didn’t become universally used until the Edo Era. Before this, the lack of voiced marks meant that there was much greater potential for puns and plays-on-words. In Modern Japanese the letter は (ha) is visually distinctive from ば (ba) and ぱ (pa). A modern Japanese author could use the letter は to mean 歯 (ha, “tooth”), 葉 (ha, “leaf”), or even 破 (ha, “middle section of a song in Noh theater”), but a Heian Era author could also use は to mean 場 (ba, “place”) or simply ば (ba, “because” [archaic usage]).
Although Japanese is astoundingly difficult for non-native speakers to wrap their heads around, the rich tapestry of sounds and characters enable authors to exercise great creativity with their choice of words and sounds. It is admirable that the Japanese people were able to take a concept hitherto unknown to them, writing, and make it their own in such a wonderfully unique way. Japanese writing would undoubtedly have been improved by a lesser reliance on Chinese influence through the centuries and a firm commitment to consistency, but despite this over-complication the system is still mostly logical in its approach to representing sounds, and has charm all its own.
Habein, Yaeko Sato. The History of the Japanese Written Language. University of Tokyo Press. 1984. Print.
Miyake, Marc Hideo. Old Japanese: A Phonetic Reconstruction. RoutledgeCurzon. 2003. Print.
Seeley, Christopher. A History of Writing in Japan. University of Hawaii Press. 1991. Print.
 Habein. Pg. 3.
 Seeley. Pg. 3.
 Miyake. Pg. 6.
 Miyake. Pg. 12.
 Habein. Pg. 9.
 Habein. Pg. 55.
 Miyake. Pg. 17.
 Miyake. Pg. 28.
 Miyake. Pg. 18.
 Habein. Pg. 13.
 Habein. Pg. 15.
 Habein. Pg. 55.
 Miyake. Pg. 43.
 Habein. Pg. 18.
 Miyake. Pg. 104.
 Habein. Pg. 57.
 Habein. Pg. 103.
 Miyake. Pg. 21.
 Habein. Pg. 22.
 Habein. Pg. 22
 Habein. Pg. 22
 Habein. Pg. 47.
 Habein. Pg. 39.
 Habein. Pg. 38.
 Habein. Pg. 25.
 Habein. Pg. 24.
 Habein. Pg. 25.
 Habein. Pg. 36.
 Habein. Pg. 34.
 Miyake. Pg. 74.
 Habein. Pg. 63.
 Habein. Pg. 28.
 Habein. Pg. 65.