核心南島語假設之辯證

In Defense of Nuclear Austronesian (and Against Tsouic)

Ross (2009) 曾提出「核心南島語假設」，在此假設下，卑南語、鄒語以及魯凱語各自為南島語的第一支分群，而所有其他的南島語則同屬於另一個第一支分群，稱之為「核心南島語」。屬於「核心南島語」這一分群的語言有一個複雜的共同創新，即在古南島語作為名物化的一些標記，在這些語言變成動詞的標記。本文分為兩部分。第一部分探討歷史語言學家在做分群時所使用的不同種類證據，並檢視支持「核心南島語假設」的證據及其他近年關於南島語分群假設的證據，依此評估這些不同假設的可信度。第二部分則討論與「核心南島語假設」不相容的「鄒語群假設」，作者認為儘管「鄒語群假設」為許多人所接受，然而並沒有證據支持此一假設。相反地，所謂「鄒語群」所反映的是長久以來鄒語和卡那卡那富語及沙阿魯阿語因為接觸而起的現象。結論認為，台灣南部山區也許就是南島語的原鄉。

Ross (2009) proposed the Nuclear Austronesian hypothesis, whereby Puyuma, Tsou and Rukai are each single-member first-order subgroups of Austronesian and all other Austronesian languages belong to a Nuclear Austronesian subgroup. The basis of this subgrouping is a complex innovation whereby certain Proto Austronesian nominalizers came also to mark indicative verbs. This paper falls into two parts. The first surveys kinds of evidence that historical linguists use in subgrouping and proposes metrics (§2) that are then applied to the innovations that support Nuclear Austronesian (§3) and other recent first-order subgroupings of Austronesian (principally Formosan) languages (§4). The second part argues that the commonly accepted Tsouic subgroup, which is incompatible with the Nuclear Austronesian hypothesis, is not supported by the evidence. Instead it reflects longterm contact between Tsou on one hand and Kanakanavu and Saaroa on the other (§5). In conclusion, it is tentatively suggested that the southern part of the Taiwan highlands appears to be the oldest Austronesian homeland area.