ABSTRACT
The data underlying Figure 7.1 is known as the so-called Kasai Data in Japanese sociolinguistics (Kokuritsu Kokugo Kenkyūjo 1966-1974; Inoue 2008a). This data shows the general tendencies of standardization of Japanese based on the lexical items of the Linguistic Atlas of Japan. The data from informants, who were born around 1895, showed an average common language usage rate of 37 percent (Kasai 1981). Data on common language use of junior high school students, who were born around 1985, were collected for the same linguistic items in a later survey (Inoue 1997). This data revealed an average common language usage of 76 percent. That is to say, a comparative analysis between these two data sets revealed an increase of 39 percent in common language use. In other words, the rate of common language use almost doubled in about 90 years (or three generations) of apparent-time difference. When drawing a straight line between the rates of these two surveys and extending it until it reaches 0 percent and 100 percent, respectively, we arrive at a time period of 230 years for the entire standardization project. That is to say, according to this linear model, people born around 1805 seem to have begun adopting common language, and the propagation of common language will be completed by persons born around 2035. People born at the turn of the twenty-first century will give birth to children around 2035. In a word, following this model, common language spread will be completed across Japan in the next generation. The inclination of this line (or rate of speed of the change) is in accord with other data on the language standardization process. Lines showing this specific inclination were also obtained for changes of the total standard forms of grammatical items. Moreover, when compared with the data
of the ratio of the residual words from dialect collections in the Edo period (1602-1868), the inclination is less steep (Inoue 1985). The linguistic realities are, however, more complex than such a linear model suggests. To start with, language change does not evolve in such a linear fashion. In language change theory, the so-called S-curve progression of language change is well-known (see Figure 7.2). According to the S-curve model, change does not occur at constant speed but follows the succession of slow-quick-quick-slow stages, which result in an S-curve of change. The underlying S-curve model of language change was confirmed by the results of the above-mentioned diachronic language study of Tsuruoka City as well as by a dialect-change survey conducted in the outskirts, Yamazoe, of Tsuruoka City. The validity of the S-curve was also confirmed by real-time changes (Yoneda 1997; Inoue 2000a; Inoue et al. 2009). As language change traces the various stages of the S-curve model, we encounter changing attitudes towards language change. These are shown in the five circles above the graph. When language change occurs and spreads, the following five stages are traversed.
