本文探討漢語序數結構「第+數詞+量詞+名詞」(簡稱為「第序數」)。首先,我們指出 Barbiers 將數字分類的特徵分析(feature analysis)不適用於漢語,因為漢語的「第序數」是規則的,不存在像 FIRST 這樣的不規則情形。此外,我們認為 Bhatt and Pancheva 將序數結構分析為最高級的論點無法解釋漢語的「第序數」。漢語的「第序數」(「第+數詞+量詞+名詞」)內含量詞,因此我們捨棄形容詞最高級分析,改採名詞結構分析,假設表序數語意的詞素「第」為定語成分(D-element),其定指性(definiteness)不明,需要根據句法與語意來確定。我們的分析不僅能解釋「第序數」的歧義,還對 H.-C. Tsai(2011)的分析進行了改良。

This paper discusses the di ordinal phrase [di + Num + Cl + N] (henceforth di ordinals) in Mandarin Chinese. We demonstrate that Barbiers' (2007) feature classification system for numerals is not applicable to Mandarin Chinese. Also, we argue that the properties of di ordinal phrases cannot be adequately accounted for by the superlative analysis proposed by Bhatt and Pancheva (2012). To address these issues, we propose a nominal analysis of Mandarin di ordinals, wherein the prefix di, denoting an ordering reading, functions as an underspecified D-element with respect to definiteness features. Our proposed analysis not only explains properties of di ordinal phrases but also provides an explanation for counterexamples found in H.-C. Tsai (2011).

序數結構; 定指; 數量; 定語結構; 特徵分析; 未指明分析

Ordinal construction; definiteness; numerals; underspecified analysis; feature analysis; DP