Paper short abstract:

The nature of hentaigana charts, or charts of non-standard hiragana characters, in Meiji textbooks is discussed. A statistical analysis of the charts shows that they largely overlap, implying that the charts differ from the characters used in the contents, contrary to earlier views.

Paper long abstract:

This paper reconsiders the nature of hentaigana charts in Meiji textbooks. Hentaigana refers to non-standard and obsolete hiragana graphs that were used before the 1900 hiragana standardisation, one of the reformations that occurred when the national compulsory education system was established. Before the standardisation, hentaigana was taught at elementary schools; hentaigana charts were one of the devices employed in education and presented in the early years of primary education. The self-contained charts have been considered to be suitable tools for observing the shifts in standardisation. A hentaigana chart is considered to present or relate to the degree to which hentaigana is used in a textbook. By examining the charts of government-edited textbooks, earlier studies made observations of the changes in the numbers of characters used in the charts and concluded that the number of characters used had gradually decreased from the beginning of the Meiji period. Although this view seems to be plausible, the basic assumption about the charts has not been proved. In addition, before nationalisation of textbooks, a larger number of textbooks were edited by the private sector, which were previously not taken into account. As a preliminary investigation, the author examined the relationship between the characters used in the contents and those used in the charts in several earlier textbooks. The result showed no clear relationship between the contents and the charts, which implies the independence of charts in regard to contents. Since the materials used in this analysis are partial, a larger and more balanced corpus is needed to confirm the results. In the present paper, the author attempts to confirm the preliminary findings and also examines the relationship of charts between textbooks using a statistical analysis of used characters compared to a larger group of charts and controlling for publisher and publishing dates. The results confirm the preliminary results and show that most of the characters in the charts overlap. The results imply that hentaigana teaching has not drastically changed during the era; in other words, the standardisation, contrary to expectations, has made radical changes to hiragana.