國立清華大學語言學研究所 新 117 次專題演講
Colloquium New Series No. 117, Institute of Linguistics, National Tsing Hua University
講者 Speaker：Alexander Adelaar 教授（University of Melbourne）
講題 Title：Favorlang: the result of language shift?
時間 Time：2026 年 5 月 13 日（星期三），中午 12:30 ─ 下午 2:00 [12:30–14:00, May 13, 2026]
地點 Venue：人社院 B305 研討室 [B305, CHSS]

Abstract
The four Western Plains Languages in western Taiwan are Babuza, Taokas, Papora and Hoanya. They represent one of the nine Formosan primary language groups distinguished by Robert Blust (1999). They ceased to be spoken in the 1960s and are currently among the most poorly documented indigenous languages in Taiwan. For most of them there are only short wordlists available, some of which are made up of less than one hundred items. The only exception is Babuza as it has a historiolect called Favorlang, for which there are various texts and a 350-page dictionary available which were published in the mid-17th century by Dutch ministers.

Several attempts have been made at the historical study of the Western Plains languages, and some scholars (Ogawa, Dyen, Tsuchida, Li, Blust) have made concerted eWorts to make the Favorlang materials available for further sourcing. However, they often did so in the context of a broader Austronesian phonological and lexical reconstruction, and it is fair to say that the Favorlang language data at our disposal have never been fully sourced for a historical study of the Western Plains languages in particular.

In this talk I will attempt to fill some of this gap. I will concentrate on several Favorlang features that seem odd in the context of Austronesian historical linguistics, such as the absence of Austronesian personal pronouns, a dearth of Austronesian vocabulary in kinship terminology, the use of the numeral 'three' for clusivity marking, the use of a mixed decimal/vigesimal counting system, and the order of prefixes. Phonologically, it often happens that several consonants reflect a single Austronesian proto-phoneme. These features are much less on display in Babuza and the other Western Plains languages. While some of these features could also be explained otherwise, their configuration suggests a non-Austronesian substrate and indicates that Favorlang may be the result of a language shift. Consequently, its linguistic history seems far more important for Formosan historical linguistics than is justified by the scant attention the language has received so far.