Cognitive Unit in Chinese Text Comprehension: Single Word or Multiple Words
|關鍵字:||中文閱讀理解;意義生成;閱讀單位;概念結合;建構整合模型;reading comprehension;meaning generation;processing unit;conceptual combinatio;construction-integration theory|
|摘要:||潛藏語意分析（Latent Semantic Analysis，LSA）在進行中文文章理解時，該以單字或是多字詞做為輸入單位是一個尚待釐清的問題。本研究以文章閱讀理解的建構－整合模型為架構，假設單字為中文文章閱讀理解的基本單位，雙字詞的詞義為單字與單字間建構－整合形成的意義，透過概念結合的機制產生雙字詞的詞義。本研究以內容分析檢驗國小一到六年級的國語課文文本中的中文單字以及多字詞中的單字是否有獨立字義，以及探討雙字詞構詞的語意原則。結果發現，99.17%的中文單字具有獨立字義，99.73%雙字詞中的單字也具有獨立字義，雙字詞構詞的語意原則依意義產生的方式可分為四大類，第一類佔20.43%，可分為屬性交集、屬性聯集、屬性互斥、類比詮釋、縮語、雙重否定。第二類佔67.79%，分別有屬性轉移、備餘、疊字、同義。第三類為關係詮釋佔8.49%。第四類只佔3.29%，可分為無單獨字義、外來語、專有名詞、狀聲詞。一、二、三類的詞義可由雙字詞的單字建構整合而得到，只有第四類的單字與詞義無關，但這類詞義多屬事物的專有名詞。因此我們推論可以透過單字與單字間意義產生的機制來形成中文雙字詞的詞義，而中文閱讀的認知處理單位即以單字作為基本單位即可。|
The purpose of this study was to investigate whether single word (character) can be the smallest unit in Chinese text processing, and how two Chinese words can be combined conceptually to produce the phrase-level meaning. The corpus of the analysis was texts taking from Taiwan’s grade1-6 Chinese textbooks. The meanings of 2662 word were checked by dictionary. The 6193 nonredundant two-word phrases were obtained after parsing and the meaning of their constituent words were analyzed. Conceptual combination theory was used to analyze the sense-making mechanism that generate the higher level meaning from joining the separate meanings of two words. Results show that 99.17% of Chinese single words have their own meanings, and 99.73% of constituent words of a two word compounds also have their independent meanings. The meanings of two-word compounds were generated through four conceptual combination mechanisms. For the first meaning generation mechanism, meaning is generated by making sense of why these two individual words were put together (20.43%) such as 聰明、明白. For the second meaning generation mechanism, meaning comes from one of its constituent words (67.79%), while another word serves as a modifier, or a reinforcer of the meaning of its head noun（美女、方正）. For the third meaning generation mechanism (8.49% ),the meaning of two-word compounds comes from the relations predicated by the two words (走路、動手). The meaning of the rest of 3.29% of two-word compunds (3.29%) is generated by arbitrarily assignment of meanings to the two-word compounds. For these compounds, such as proper names, loanwords, or onomatopoeia, phrase-level meanings have nothing to do with the meanings of its constituent words. Only 7 out of 6193 two-word Chinese compounds cannot be explained by these four types of sense-generation mechanisms. It can be concluded that the meanings of Chinese multiple words can be generated through integration of the meanings of their constituent words, and single word itself can be the basic cognitive unit in Chinese reading comprehension.
|Appears in Collections:||Thesis|