兩種量詞語言類型：排灣語名詞類別語素的類型研究

Two Types of Classifier Languages: A Typological Study of Classification Markers in Paiwan Noun Phrases

數詞與量詞的共存限制以及複數詞與量詞的共存限制一直都是語言學研究的重要課題，其中更有不少文獻致力於詮釋漢語與英語在這兩種限制的相同點與相異點。這篇文章除了研究排灣語、其他台灣南島語以及藏緬語的相關詞彙、句法和語意行為之外，並說明Cheng 和Sybesma (1999)、Li (1999)和Chierchia (1998) 等人的主張都無法有系統的處理英語、漢語、台灣南島語以及藏緬語的相關語言類型異同。文中因而建議：除了要考慮數詞、量詞和複數詞的不同詞彙、句法和語意結構之外，還要將所謂的量詞語言細分為貧脊型的量詞語言和豐富型的量詞語言。

Recently many claims have been made concerning the contrasts in the occurrence of classifiers in the presence of numerals with nouns as well as the nonoccurrence of plural morphology in the presence of classifiers with nouns between languages like Chinese and those like English. This paper examines some relevant analyses such as Cheng & Sybesma (1999), Li (1999), and Chierchia (1998) in view of the morphological, syntactic, and semantic behavior of numerals in languages like Paiwan. Our findings in Paiwan, together with those in other Formosan and TibetoBurman languages, indicate that these accounts are all problematic in that they cannot capture the cross-linguistic distribution of numerals with classifiers, plural morphology, and nouns. For a more plausible approach, it is suggested that, in addition to the cross-linguistic variations in the morphological, syntactic, and semantic structures of numerals, classifiers, and plural morphology, so-called classifier languages should be further distinguished as poor- or rich-classifier languages.