本文討論漢語中量詞後加“的＂這一語法現象。我們認為，漢語的量詞，比如“瓶＂，有計數和計量這兩種功能，但是“的＂的出現，我們只能獲得計量這一功能。我們同意 Tang (2005) 和 Hsieh (2008) 的觀點，認為“的＂可以出現在個體量詞後面，但是該量詞短語的數詞必須是約整數，如百、千、萬等或者有“大概＂、“左右＂等表示約量的副詞的修飾。從語義上來看，這是一種“近似化＂，是一種計量功能，在這種解讀下，個體量詞不再具有本來的計數功能，相反是獲得了計量功能（計量短語中的中心語）。在 X.P. Li (2011) 觀點的基礎上，我們認為，數–量–名短語在計量解讀時的結構為 [數–量[名]]，當“的＂出現的時候，它修飾的是 [數–量] 這一成分。這一處理使得我們能對結構助詞“的＂有一個統一的結構和語義。

In this paper, we show that the particle de always induces a measure reading when it follows a classifier head. We show that classifiers such as ping ‘bottle’ are ambiguous between a measure and an individuation or counting interpretation, but that de always induces the measure reading. We follow Tang (2005) and Hsieh (2008) in arguing that the particle de is possible after sortal or individuating classifiers which occur with a high round number or an approximative marker, but we argue that this is because approximation is semantically a kind of measure function, and that in these contexts the sortal classifiers are interpreted as heading a measure phrase. We claim further, following X.P. Li (2011) that when a numberclassifier sequence is interpreted as measure phrase, the string “Num-Cl(-de-)N” always has the syntactic structure [[[Num-Cl](-de-)]N], with Num-Cl forming a constituent which de takes as a complement. This allows us to suggest that de in classifier constructions has the same grammatical function as in other modificational contexts, and paves the way for a unified analysis of de. We conclude with an explicit semantics for the interpretation of [[[Num-Cl](-de-)]N] strings.