Late Old Japanese (LOJ), the language of the Heian period and therefore essentially ‘Classical Japanese’, was pronounced differently from earlier and later Japanese. Whereas Early Old Japanese had eighty-eight different possible syllables, the native vocabulary of LOJ had sixty-six (or sixty-seven, depending on whether one believes ye and e were distinct), plus two ‘syllabic consonants’ m and n. Kana as a script emerged at the beginning of the Heian period, and fairly accurately reflects the contemporary system. However, originally the letter(s) for mu were used for the ‘syllabic consonants’ m and n, and ye and e (which many linguists believe were still distinct in the Heian period) were written with the same letter(s). For this reason we normally write ‘mu’ instead of ‘m’ or ‘e’ instead of ‘ye’.
In terms of pronunciation, it’s probable that t, k, m, n, r, w, y were pronounced the same as in Modern Japanese. The ‘prenasalised’ sounds b, d, g, z were similar to Modern Japanese, but had a sort of nasal effect at the beginning. Imagine the type of ‘b’, ‘d’ or ‘g’ that you pronounce when you have a badly blocked up nose, and that’s exactly how EOJ, LOJ and MJ pronounced b, d, g. F was probably pronounced like f in Modern Japanese. P was a borderline sound, probably occurring only in onomatopoeic expressions or in interjections.
LOJ had just five vowels: a, e, i, o, u. They were probably pronounced with the same pronunciation as a, e, i, o, u in Spanish. This means that a, e, i and o were pronounced as in Modern Japanese, but u probably involved more rounded lips than in Modern Japanese.
LOJ had tone-accent, just as Modern Japanese does. The features of tone-accent in LOJ parallel those of tone-accent in Modern Japanese: most vowels were either pronounced with high pitch (H) or low pitch (L); only one stretch of high pitch was allowed in a word; nouns represented quite a variety of pitch patterns over the word, so it seems that there were five different accentual possibilities on a two-syllable noun; and verbs and adjectives, on the other hand, allowed only two patterns.
Apart from the eleventh-century Ruiji Myôgishô dictionary, tone-accent wasn’t written and so is not represented in the transcription. What Ruiji Myôgishô shows us, though, was that most inflectional endings on verbs, including the te-form, had their own accentual pattern separate from that of the preceding verb, suggesting that these were actually pronounced as separate words: i.e. uta yomite ‘he composed a poem, and’ or uta yomitari ‘he (has) composed a poem’ were probably really pronounced uta yomi te and uta yomi tari. Linguists normally still transcribe these as yomite and yomitari under the influence of Modern Japanese usage.
Just as court society was becoming heavily influenced by Chinese society, the formal language was also becoming influenced by Chinese. Aristocratic men were expected to be proficient in writing and reading Literary Chinese, although the norm was to read Chinese out loud by translating it into Japanese. Even though very few men outside of the Buddhist priesthood are likely to have actually spoken Chinese, they also frequently composed classical Chinese poetry! This meant they had to have some idea of how Chinese was pronounced, because, unlike Japanese poetry, Chinese poetry rhymed and had a strict system of metre. In the areas of government, the military, court ranks and Chinese culture, loads of Chinese loan-words were used by the aristocracy.
Chinese loan-words introduced sequences of sounds that were alien to Japanese:
- Several combinations of consonants were found in Chinese loan-words: kw, gw, ky, gy, ty, dy, ny, ry, sy, zy, fy, by, my, and even briefly wy. (It’s possible that the kk, tt, ss, pp that occurred in later Japanese also existed at this stage, but this is by no means clear.)
- Several combinations of vowels were found in Chinese loan-words: au, eu, iu, ou, uu, ai, ei, ui.
- Two consonants, n and m, could occur at the end of syllables or words.
Although Chinese words are exceptionally rare in LOJ poetry, which preferred to maintain linguistic purity, the fact that they were used in spoken Japanese meant that they influenced the way native Japanese words were pronounced. Whereas in EOJ there had been a virtual prohibition on two vowels in a row, the fact that many Chinese loan-words had two-vowel combinations meant that it no longer sounded so strange, and so increasingly even native words began allowing two-vowel combinations. In the middle of words f became w, and w in turn was often dropped except before a. Thus, LOJ kefu ‘today’ may have been pronounced as ‘keu’, kinofu ‘yesterday’ as ‘kinou’, kafi ‘shellfish’ as ‘kai’. This was to have an immense impact on the pronunciation of later Japanese.
We know that these changes had happened, because manuscripts start writing wo for fo or o, o for wo or fo, and fo for wo or o. This shows that the syllables represented by these three spellings were no longer distinct from each other. It’s normal, though, to ignore the spelling inconsistencies and to transcribe LOJ according to an idealised pronunciation where o, wo and fo were distinct.
If you have any questions about Late Old Japanese phonology, please email Dr. Nic Tranter.