Jodai Tokushu Kanazukai (ancient Japanese phonetic orthography) (上代特殊仮名遣)

Jodai Tokushu Kanazukai is Japanese phonetic orthography used in ancient Japan (around the Nara period), which can be observed in the documents written in Manyo-gana (Chinese characters used as phonetic characters) such as "Kojiki" (The Records of Ancient Matters), "Nihon Shoki" (Chronicles of Japan) and "Manyoshu" (Collection of Ten Thousand Leaves). It is recognized as a historical phonetic orthography in ancient times, preceding Teika Kanazukai (phonetic orthography established by FUJIWARA no Teika in the Kamakura period). Its name has been derived from the article called 'Special Phonetic Orthography and Usage in the Documents Written in Ancient Japanese' written by a Japanese linguist Shinkichi HASHIMOTO. It is also called simply "Jodai-gana."

It is known that among the 50-character syllabary used today, 14 sounds were written in two types of Manyo-gana, namely A-type and B-type, in ancient Japan before the Nara period, and this distinction of the types was strictly maintained; these 14 sounds were "キ, ヒ, ミ" (ki, hi, mi) with the vowel of イ (i), "ケ, へ, メ" (ke, he, me) with the vowel of エ (e), "コ, ソ, ト, ノ, (モ), ヨ, ロ" (ko, so, to, no, [mo], yo, ro) with the vowel of オ (o), and the vowel "エ" (e). However, the distinction of モ (mo) is observed only in "Kojiki." "エ" appeared in two sounds: one of vowel (e) and the other of a consonant and vowel (ye). The distinction between A-type and B-type also existed in the dull sounds of ギ, ビ, ゲ, ベ, ゴ, ゾ and ド (gi, bi, ge, be, go, zo and do).

As for the distinction between A-type and B-type, for example, Manyo-gana representing the sound of "き" (ki) included in the Chinese characters of 支, 吉, 峡, 来 and 棄 were A-type characters and used to write "き" (ki) of "aki" (autumn), "kimi" (you), "toki" (time) or "kiku" (listen). On the other hand, characters like 己, 紀, 記, 忌 and 氣 were B-type, and used to write "き" (ki) of "kiri" (mist), "kishi" (shore), "tsuki" (moon) or "ki" (tree). We can see such a methodical usage of phonetic characters throughout the documents written in ancient Japanese, and with some exceptions, A-type and B-type were strictly distinguished and there was no confusion between them.

A theory has been developed about this distinction between A-type and B-type, saying the distinction stemed from sounds; in ancient times, there were eight vowels, which were five vowels of アイウエオ (a, i, u, e, o) with each of イエオ (i, e, o) having two types, A-type and B-type. Thus, it has been considered that ancient Japanese syllabary had 87 sounds (or 88), not 50. It says that the distinction of such types gradually faded away after the Heian period and they were unified. However, actual sounds for characters are largely unknown, and there are also different opinions whether they had different phonemes or not.

At the time Jodai Tokushu Kanazukai was used, "Kana" (Hiragana and Katakana, phonetic Japanese syllabaries) had not been developed yet so that only Kana character representing Jodai Tokushu Kanazukai is エ with the consonant y (ye). Therefore, when it is necessary to distinguish between A-type and B-type, characters are written with underlines, in Katakana or alphabet letters with umlauts.

Discovery of Jodai Tokushu Kanazukai
Norinaga MOTOORI and Tatsumaro ISHIZUKA
The study of Jodai Tokushu Kanazukai was initiated by Norinaga MOTOORI.
In "Kojiki-den" (Commentary on the Kojiki), Norinaga's extensive commentary book on "Kojiki," he already pointed out in the section 'matters concerning Kana' of the first volume, that 'even for the same sound, it uses different phonetic characters depending on the words.'
However, it was only a limited comment, and this idea of Motoori was further developed by his disciple Tatsumaro ISHIZUKA who wrote "Kanazukai Oku no Yamamichi" (The Mountain Road into the Secrets of Kana Usage) (published around 1798). In this book investigation was carried out on the usage of characters in "Kojiki," "Nihon Shoki" and "Manyoshu" in which Manyo-gana was used. Ishizuka, in this book, concluded that there was a distinction in Manyo-gana, of the types of characters used for fifteen sounds, エ, キ, ケ, コ, ソ, ト, ノ, ヒ, ヘ, ミ, メ, ヨ, ロ, チ and モ (e, ki, ke, ko, so, to, no, hi, he, mi, me, yo, ro, chi, and mo). However, the texts used for the investigation contained some errors because there was not much criticism on the credibility of the materials at that time. Furthermore it did not reach any conclusion to connect the distinction of the types of phonetic characters with the difference of phonemes, and thus did not draw much attention.

Theory of Eight Vowels
Although the studies by Norinaga and Ishizuka had not been evaluated for long time, Shinkichi HASHIMOTO rediscovered them and published an article called 'A Discovery in the Field of Japanese Kana Usage Research Concerning Kanazukai Oku no Yamamichi by Tatsumaro Ishizuka' for 'Teikoku Bungaku' (Literature of the Empire) and it made the studies recognized by the academic community.
The studies conducted after Hashimoto have not accepted the distinction of the types of characters used for "チ" (chi), the distinction which was pointed out by Tatsumaro ISHIZUKA
It has been considered that in ancient times there was a special distinction of the types of characters used for 14 sounds of エ, キ, ケ, コ, ソ, ト, ノ, ヒ, ヘ, ミ, メ, ヨ, ロ and モ (e, ki, ke, ko, so, to, no, hi, he, mi, me, yo, ro and mo) and their dull sounds if any. Hashimoto called this distinction of the types of characters "Jodai Tokushu Kanazukai." As for モ (mo), such distinction is only observed in Kojiki. As for the difference of sounds between A-type and B-type, Hashimoto presumed that A-type characters had the sounds of i, e, and o, same as the current ones, while B-type had the sounds of ï, ë and ö.

Hideyo ARISAKA and 'Arisaka's Law'
Hideyo ARISAKA published in 1934 an article called 'The Laws of the Combination of Syllables in Ancient Japanese', in which he pointed out the following laws concerning Jodai Tokushu Kanazukai.

A-type syllables with the vowel of オ (o) and B-type syllables with the vowel of オ (o) are not combined into one unit.

Syllables with the vowel of ウ (u) and B-type syllables with the vowel of オ (o) are rarely combined into one unit. Particularly, when the unit consists of two syllables, they are never combined.

Syllables with the vowel of ア (a) and B-type syllables with the vowel of オ (o) are rarely combined into one unit.

In fact these laws had already been mentioned in 1932 in his article 'Concerning the Usage of Kana for mo in Kojiki,' but he pointed them out with firm conviction in the aforementioned article. It was called Arisaka's Laws. This tendency that vowels were divided into groups and often combined with other vowels of the same group and rarely with the ones of other groups was considered as a trace of 'vowel harmony' observed in Turkish for example. Arisaka's Laws were also considered as evidence for Japanese belonging to an Altaic language family.

Jodai Tokushu Kanazukai became an established theory in Japanese linguistics by these studies of Hashimoto and Arisaka. They criticized Jindai-moji (written language claimed to have existed before the introduction of Chinese characters) used in the so-called Koshi Koden (unofficial ancient documents) including Takeuchi-monjo (alleged record of the lineage of ancient gods and a dynasty preceding Emperor Jinmu) for 'having only five vowels and not adopting the ancient Kana usage of eight vowels' and denied it as having been forged in a later era when such distinction of Kana usage had disappeared. Concerning this dispute, in 1942, Hashimoto appeared as a witness for the prosecution in the trial of Takeuchi-monjo, the second oppression against Amatsukyo sect.

Criticisms on the Theory of Eight Vowels
The theory that ancient Japanese had eight vowels had been widely accepted and had become almost an established theory, however after 1975, several opposing theories appeared. The first one was 'A Study on Ancient Japanese Vowel Structure - Attempt for Internal Reconstruction' by Katsumi MATSUMOTO. The internal reconstruction is an approach to the history of certain language through not only comparison with other languages but also a study on the synchrony of the language.

Matsumoto criticized Arisaka's Laws for the combination of syllables that the idea of 'combined into one unit' was ambiguous, and insisted that the laws for the combination should be investigated in relation to all vowels, not only the vowels that had the distinction between A-type and B-type. Then he carried out the investigation, grouping the vowels into three types based on the study of Yoshisuke FUKUDA in 1965. As a result, he concluded that the vowels that were previously considered to have the distinction between A-type and B-type indicated complementary distribution and thus these vowels were not distinguished from each other but their phonemes were identical. Matsumoto took Greek letters for/k/ for example and explained that the existence of two types of letters, k and q, for/k/ did not necessarily indicate the fact that these two consonants were distinguished when used. Similarly, he pointed out that the distinction of usage in Jodai Tokushu Kanazukai did not precisely represent the vowel system of the time. Furthermore, Matsumoto argued that Japanese vowels had developed as follows.

Three vowels of i, a, u

Four vowels of i, a, o, u

Six vowels of i, e, Ï, a, o, u

Current five vowels

He then concluded that ancient Japanese vowel system had five vowels same as the current system.

The article by Katsumi MATSUMOTO was published in March 1975, and in September of the same year, Satoshi MORISHIGE almost coincidentally published 'What is Jodai Tokushu Kanazukai' and criticized the theory of eight vowels of Jodai Tokushu Kanazukai from a different point of view from Matsumoto's. It was published in September but written in February of the same year, and thus the criticisms on the 'established theory', the theory of eight vowels, were written almost in the same period.

Morishige firstly explained that when indeclinable words were used as exclamation without particles, the sound of "イ" (i) might be added to the words as a substitute for particles. It meant that the exclamation of "Hana" (flower) might not be expressed as "Hana-yo" (with a particle) but "Hanai," "Haina" or "Hainai" by adding "イ" (i) to the vowels. The summary of Morishige's theory was that B-type sounds with the vowels of エ (e), イ (i) and オ (o) were formed by adding イ (i) to the sounds with the vowels of ア (a), ウ (u) and オ (o) respectively.

Morishige also concluded that Japanese vowel system had been consisted of five vowels. According to his theory, the distinction of characters observed in Manyo-gana was caused by Toraijin (people from overseas, especially from China and Korea, who settled in early Japan and introduced Continental culture) who had distinguished the difference of the sounds that were not unimportant for Japanese. Morishige explained it taking the Hepburn system of romanization for example, which distinguished the sounds not necessarily important for Japanese such as sh, ch, ts or f. He also pointed out as evidence that although only the sound of "コ" (ko) among Jodai Tokushu Kanazukai had survived until the beginning of the Heian period, no such distinction had existed in Hiragana.

Arguments over New Theories
Many criticisms developed against these new theories published in the same period, and from December 1970 (?) to November of the following year, the discussions about the theories were held four times in Mainichi Newspapers. Particularly discussions held in a magazine "Gengo" (Language) were intense. The June issue of 1976 featured 'investigation of vowel harmony,' with Shiro HATTORI's article 'Ancient Japanese Vowel System and Harmony' that advocated for the theory of six vowels and Matsumoto's 'Japanese Vowel Structure' appearing next to and criticizing each other. Further, in the August issue Susumu ONO introduced both theories and presented his own view in 'Old Japanese Vowel System,' which was, in turn, countered by Matsumoto, in 'Concerning A-type and B-type Manyo-gana Characters with the Vowel of オ (o)' in the November issue, and by Hattori, saying 'Old Japanese Vowel Phonemes were Six, not Eight' in the December issue.

In 1981 Hiromichi MORI, in his 'The Northern Dialect of the Tang Dynasty and the Phonetic Value of the Vowels in Old Japanese', published a theory of seven vowels, recognizing the distinction between the A- and B-types other than those with the vowel of エ (e), and tried to presume concrete sounds using the northern dialect of the Tang dynasty and phonetically arranged dictionary of ancient Chinese characters.
In response to it, Hisao HIRAYAMA had discussions with him in a magazine "Kokugogaku" (Studies in Japanese Linguistics) through a series of articles called 'Concerning Mr. Hiromichi Mori's Theory that Nihon Shoki α Group Was Written Based on Original Chinese Sounds,' 'In Response to Mr. Hisao Hirayama, Further Argument for the Theory that Nihon Shoki α Group Was Written Based on Original Chinese Sounds' and 'Discussing Again Concerning Mr. Hiromichi Mori's Theory that Nihon Shoki α Group Was Written Based on Original Chinese Sounds.'

Recently no study has been published in which the previous theories were denied fundamentally, but it does not necessarily mean we have reached conclusion concerning ancient Japanese vowels. In the future approaches from various fields are expected concerning the matters like how many vowels actually existed in ancient Japanese or why only a part of characters with the vowels of イ (i), エ (e) and オ (o) had the distinction of the types.