Study the striking characteristics of Japanese pronunciation
Japanese language institutes in Delhi teach you a foreign language in quite systematic way. Let us see the case of Japanese pronunciation. A striking characteristic of Japanese pronunciation is the frequent occurrence of voiceless vowels. These result from the application of a vowel unvoicing process which is conditioned by a number of factors the most important of which is that the consonants surrounding the vowel (or preceding it if the vowel is followed by #) must be voiceless. In Japanese such consonants are limited to the three stops [p, t, k], the two affricates [ts, ts?] and the five fricatives [s, s?,?, ç, h] e.g. /suki/? [su?ki] 'likeable'; /sita/? [s?i?ta] 'below' ; /hito/? [çi?to] 'man'; /huta/? [?u?ta] 'cover'; /tuku/? [tsu?ku] 'to attach' ; /titi/? [ts?i?ts?i] 'father'; /pika/? [pi?ka] 'flash' ; /kusi/? [ku?s?i] 'comb'; /kita/? [ki?ta] 'came'; and /desu/? [desu?] 'is'.
Another important condition is the sonority of the vowel, the less sonorous high vowels being more likely to unvoice than the more sonorous lower ones e.g. the unvoiced non-high vowels in [ko?koro] 'heart'; [to?koro] 'place'; [ke?kkoo?] 'all right'; [ata?takai] 'warm'; [ha?ha] 'mother'; and [waka?katta] 'understood' occur under more restricted conditions than those governing the unvoicing of the high vowels in above mentioned examples. Example faster rate of speech and according to Martin (1952 p. 14) when unvoicing of non-high vowels occurs 'it is usually of the, initial syllable and coinciding with the repetition of the same vowel in the following syllable...' Identity of the preceding and following consonant would also seem to be a contributory factor.
In a study of vowel unvoicing based on spectrographic data, Ban (1962) proposed inherent vowel length as a conditioning factor in vowel unvoicing. She found that other things being equal the inherent length of the five Japanese vowels differed non-distinctively as follows with the shortest vowel /u/ representing a ratio of one: /u/ /i/ /o/ /e/ /a/
The correlation between vowel height, inherent length, and sonority suggests that a single feature may be involved here. Other significant variables studied by Han were: the effect of tempo (the faster the tempo the more unvoicing), pitch accent (low-pitched vowels are more likely to unvoice than high-pitched ones) and the manner of articulation of neighboring sounds (fricatives, affricates, and stops in that order are more likely to cause unvoicing). Thus she found that in comparing ki?kukotodesu 'is to listen'; and kiku?kotodesu 'is with Kikuko' it is the first vowel in first example that is unvoiced while in another, it is the second. The difference is due to the difference in pitch level. On the other hand in si?kutoitta 'he said "four by nine"' the first vowel is unvoiced regardless of the high pitch, showing the overriding effect of the fricative manner of articulation.