Universals of tone rules and diachronic change in Japanese

In: Journal of Asian and African Studies, 94 (2017), pp. 217-242. ISSN 0021-9096

Preview

PDF, English Download (577kB) | Terms of use Download (577kB)

Abstract

The paper examines two competing theories on the historical development of Japanese tone (Kindaichi's theory or the standard theory and Ramsey's theory), in light of universals of tone rules. It is argued that many developments that must have occurred widely, according to the standard theory, are unnatural when compared to developments in other languages.

Rightward tone shift plays an important role in the standard theory, and the naturalness of rightward tone shift is beyond doubt. However, many other developments that form part of the standard theory are problematic, as they involve complicated changes for which no phonological motivation can be offered.

The alternative theory, and its reconstruction of the Middle Japanese tones, explains the wide geographical distribution of the Tokyo type tone systems, as this type is closest to proto- Japanese. Two relatively simple changes (assimilation of the tone of the particles in some dialects, followed by a tone reduction in all dialects) account for the different merger patterns in the modern Tokyo type dialects. Furthermore, the well-established fact that long vowels offer stronger support for contour tones than short vowels, explains the geographical distribution of the Nairin type versus the other types.

The radical tone reduction after the Middle Japanese period is the historical change in Japanese tone that was most far-reaching, and the polarizing effect of a HL interval explains why only H followed by L in Middle Japanese was preserved in the modern dialects. This phonological explanation may also account for the very similar tonal development known in Bantu linguistics as 'Reverse Meeussen's Rule', for which no phonological motivation has been proposed so far.

The leftward shift of H tone, which according to Ramsey's theory, took place in the Kyoto type dialects, has parallels in Bantu languages that went through a similar tone restriction as Japanese. Bantu languages tend to be rich in productive prefixes, so that word-initial H tones, when shifted, will be realized on the prefix. In a predominantly suffixing language like Japanese, leftward shift of word-initial H tone had a different result, namely the development of word- initial L tone.