Some Remarks on the Linguistic Position of Thao

Abstract

Over the past two decades, much progress has been made in the study of the aboriginal languages of Taiwan. Despite these advances, many of the fifteen surviving languages remain inadequately described, and their histories poorly understood. The major purpose of this paper is to show that Thao, one of the most endangered and underdocumented languages, has been misclassified because of insufficient attention to the problem of direct and indirect inheritance. A careful consideration of the linguistic position of Thao sheds light on the broader ethnohistory of central Taiwan, and highlights the importance of fundamental principles of historical method that extend well beyond this particular case. 1. PREVIOUS CLASSIFICATIONS OF THAO. The first serious attempt to subgroup the languages of Taiwan was that of Dyen (1963, I965), who did not consider Thao. Shortly thereafter, Ferrell (1969) proposed a classification of all languages, including those known only from seventeenth century Dutch religious documents and from late nineteenth-early twentieth century Japanese scientific surveys. The result was a division into three major groupings: Atayalic (Atayal, Sediq), Tsouic (Tsou, Kanakanabu, Saaroa), and (all other languages). Ferrell further subdivided into two major branches, called I and II, but at the same time he clearly expressed his reservations about the reality of as a genetic unit. In the case of Yami, closely related to the languages of the Batanes and Babuyan Islands in the northern Philippines, it was apparent from the beginning that the term Paiwanic was more a temporary convenience than the label for a seriously defended hypothesis about linguistic history. For most scholars, Ferrell's tripartite classification of the languages has continued to serve as a convenient model until the present. The first challenge to the reality of was presented by Li (1985), who argued that the Atayalic languages should be combined with Saisiyat and Pazeh into a larger Northern Formosan group. Although this scheme did not affect the classification of Thao, Li (1990) proposed a comprehensive subgrouping of the extant languages, in which Thao and Bunun form a branch of Southern