Regional Stereotypes and the Perception of Japanese Vowel Devoicing 193devoicing in Kinki is very individual and that some Kinki speakers devoice
vowels as frequently as those in Tokyo do.
The data on devoicing in Kinki in these previous studies, however, are
taken from very limited phonological environments or social variants, and
the devoiceable vowel is not always compared with a vowel in the same pho-
nological environment; for example, if /u/ is the last sound in a word with
HL (high then low pitch) in Tokyo but LH (low then high pitch) in Kinki,
/u/ is unaccented in the ¿ rst but accented in the second. Kinki speakers may
avoid devoicing in such cases because of accentuation, just as Tokyo speak-
ers would, and such comparisons would lead to an inaccurate evaluation of
overall devoicing rates.
Table 7.1 Variation in Devoicing
Tokyo Subjects Osaka (Kinki) Subjects
Devoiced Nondevoiced Devoiced Nondevoiced
Sugito (1969) 65.8% 34.2% 29.4% 70.6%
Sugito (1988) 55.6% 44.2% 32.2% 67.8%
Fujimoto (2004) 74.8% 25.2% 56.0% 44.0%
Yo s h i o k a (19 81) 5 6. 5% 43. 5% N /A N /A
Tahara et al. (1988) N/A N/A 82.5% 17.5%Table 7.1 shows the devoicing rates in several studies. The rates by Tokyo
speakers are quite similar in Sugito (1969), Sugito (1988), and Yoshioka
(1981). The rate for Osaka (Kinki) speakers is higher in Fujimoto (2004)
than in Sugito (1969) and Sugito (1988) and even higher in Tahara et al.
(1998), but the vowels in Tahara’s data are all in the optimal devoicing envi-
ronment—a high vowel between two voiceless consonants in an unaccented
mora. Re-calculated devoicing rates in just this environment for the other
studies are shown in Table 7.2. The devoicing rate in Tahara’s data (Kinki,
Table 7.1) and the three data sets of the tokens in similarly unaccented morae
(Tokyo) in Table 7.2 are quite comparable. It will be necessary to collect a
larger amount of more controlled Kinki production data to determine the
distribution and variation of devoicing before making such comparisons.
Nevertheless, these more recent comparisons make it possible to assume
that the devoicing rate in Kinki is not all that different from that in Tokyo,
at least in the environment where devoicing is most frequent, and I take this
as a tentative assumption.