本文旨在探討詞彙知識理論中之「模組�|屬性動詞語意表達模式」（MARVS, Biq 2000），並將此理論延伸到多義動詞的分析。承續「不同的事件結構必定有語意上的區分」的理論，並用於“put”和“set”兩動詞的語料庫為本分析。我們發現必需要有更深一層次的語意區分。換句話說，若有兩種意義（meaning）可同時在一語句上存在，則我們應將此兩詞義表達視為意面（meaning facet），而非為不同之詞義（sense）。此理論乃承繼安可思等（1998）對名詞多義狀態所提的理論。總而言之，在本文中，我們利用跨語言詞彙語意之研究來支持MARVS 理論架構以及限定動詞歧義之定義。

　　　　　This paper explores the theory of lexical knowledge set forth in the Module-Attribute Representation of Verbal Semantics (Biq 2000) and extends its scope to include issues concerning verbal polysemy. Previous versions of the theory postulated that different event structures required sense distinctions. In examining corpora data for the verbs 'put' and 'set' we argue that an additional criterion for 'sense' needs to be put in place, namely the standard that if two meanings can coexist in the same sentence, then their representation is at the meaning facet, and not the sense level (following Ahrens et al. 1998). In sum, this paper adds to the cross-linguistic evidence supporting MARVS as well as constrains its definition of ambiguity.