A Reader in Sociophonetics

198 Midori Yonezawa Morris

Pitch Accent %

Tokyo PA (i.e., “nonlocal”) 86.84

Kinki PA (i.e., “local”) 87.60

Perhaps these different patterns reÀ ect the different positions and values
of local varieties in Tokyo and Kinki. The Kinki dialect functions to bond
people due to its long history in the area, while Tokyo people may well have
a smaller sense of solidarity due to the short history of its status as a standard
variety and perhaps even due to a lesser sense of local identity in an area
where many residents have immigrated from other regions of the country.
These results show a similar tendency to ones reported in such studies as
Preston (1996), in which regional varieties are valued by their speakers along
such different dimensions as “correctness” and “pleasantness.”

5.2 Voicing variants

The main results of this study, however, have to do with the perception of
devoicing. Tables 7.9 and 7.10 summarize the results for voicing variants, with
both percentages and weights obtained from Goldvarb. Again, both Tokyo and
Kinki respondents tended to judge the speaker’s region based on the stereotypi-
cal criterion of devoicing, a low-level allophonic feature. In other words, their
judgments are promoted by a feature that is only putatively their own—devoic-
ing in Tokyo and nondevoicing in Kinki. This is a weaker tendency in Kinki,
as shown by lower percentages and a smaller range of weights. This weaker
tendency for expected judgments of voicing variants in Kinki also supports
the idea that there are higher devoicing rates in Kinki speech than believed.
If Kinki people devoice vowels as frequently as in Tokyo speech, at least in
the most optimal environments, they may well not use devoicing variation
ef¿ ciently as a criterion in making judgments. It is possible that Kinki people
assume that devoicing is a non-Kinki feature based on the higher nondevoicing
rate in certain low-frequency devoicing environments (e.g., accented morae),
but con¿ rmation of that interpretation would require a different study.

Voic i ng % / We ig ht

Devoiced 70.11/0.552

Nondevoiced 59.90/0.436

Table 7.8 Kinki Results PA = Pitch Accent

Table 7.9 Tokyo Results by Voicing Variants