本論文主要採用心理語言學途徑來探討中文名詞是否像英語名詞一樣具有標記分離性之語言機制。我們假設與形狀分類詞高度共現的名詞具有[+分離性]的標記,而與形狀分類詞共現率較低的名詞則有[-分離性]的標記。我們設計了二個問卷形式的調查和三個線上實驗來探究這個議題。在數量向度評量的調查裡,我們檢驗具有高形狀分類詞共現率的名詞相較於低共現率的名詞是否會引發較多以數目為判斷依據的效應。結果並不顯著。我們認為這是因為感官認知因素介入的結果。然而,在語句完成的調查裡,我們發現高形狀分類詞共現率的名詞比低共現率的名詞更會引發受試者製造具有非泛指的語句。在自我調度的閱讀實驗裡,我們發現當具低形狀分類詞共現率的名詞出現在量化的語境下,受試者需要更多的時間來閱讀理解。而具有高形狀分類詞共現率的名詞則需要較短的閱讀時間。我們認為低形狀分類詞共現率的名詞引發的時間效應是因為其具有[-分離性]的語意,所以會在量化的語境下歷經語意誘迫的歷程。另外,在眼動閱讀實驗裡,也發現到閱讀時間、回溯機率和回溯時間上的效應。在量化的語境下,低形狀分類詞共現率的名詞會引起較長的閱讀時間、較高的回溯機率及較久的回溯時間。最後,在眼動視覺情境的實驗裡,我們也發現當受試者聽到低形狀分類詞共現率的名詞時,注視繪有單一目標物圖片的機率會比聽到高形狀分類詞共現率的名詞時較低。本研究證明了中文名詞的確具有不同的指涉特性。同時,就像英語一樣,也有句法提示機制來幫助察覺詞彙內部具有不同分離性的語意特質。
This dissertation investigated whether Chinese nouns, like English nouns, have a linguistic apparatus of individuation using a psychological approach. We hypothesized that nouns that highly collocate with shape classifiers are lexically specified with [+individual] while nouns that do not are marked with [-individual]. We conducted two off-line and three on-line experiments to pursue this issue. In the dimension rating task, we intended to examine whether nouns that habitually go with shape classifiers tended to give rise to a number-based quantity judgment, as opposed to nouns that do not. The results did not show an effect. We suggest that the failure was due to a conceptual comfound. However, in the sentence completion task, nouns that frequently take shape classifiers showed a stronger tendency to incur non-generic sentences relative to nouns that are less compatible with shape classifiers. In the self-paced reading task, the results demonstrated that a longer time was taken in sentence reading in a quantificational context when the noun being quantified is less inclined to go with shape classifiers. On the contrary, a shorter reading time was consumed when the sentence contained a noun that tends to take shape classifiers. We suggest that the effect occurred because nouns that barely take shape classifiers have non-individuated denotation and consequently a coercive operation was triggered. In the eye-movement reading procedure, a marked effect of condition emerged on reading times as well as regressive probability and duration. In a context of quantification, nouns that scantly take shape classifiers caused not only a longer total reading time but also a higer probability and a longer duration of regression. Lastly, in the eye-movement visual world study, the findings indicated that, given a noun that less frequently occurs with shape classifiers, the probalbity of fixating the picture depicting the single instance of the target object was lower in comparison to a noun that frequently collocates with shape classifiers. The findings provided evidence that Chinese nouns are distinguished in terms of reference. Furthermore, akin to English, Chinese provides a syntactic cue to individuation at the lexical level.