閱讀文字和圖片以建構空間結構和動態表徵的眼動研究

Conveying Spatial and Kinematic Representations in Text Reading via Words and Pictures: An Eye-movement Analysis

本研究探討讀者如何透過閱讀文字和圖片以建構空間結構和動態表徵，以及此兩種表徵之間的關係為何。本實驗為兩階段設計：在第一階段，大學生受試者被隨機分派到文字組（閱讀描寫儲水槽空間結構的文本）或圖片組（看一張儲水槽結構的圖片），然後，完成一個空間結構測驗；在第二階段，兩組讀者閱讀一篇相同的文章，內容描述儲水槽的動態運作歷程，並讓受試者修改他們第一次空間測驗的答案。受試者閱讀動態訊息文章時的眼動型態會被記錄下來。研究結果顯示，文字組在空間結構測驗中畫出儲水槽內各個部件（如：把手、連接桿、上下圓盤等）之間的關係較正確，代表文字較有助於讀者形成部件之間的接續關係；但圖片組畫出來的圖與原圖相似度較高，代表圖片較有助於讀者形成整體結構的空間表徵。而從第二次修改測驗的結果發現，讀者在讀完動態訊息文本之後所形成的動態表徵主要是幫助讀者修改空間表徵中的「部件接續關係」，而非「部件與整體的位置關係」。此外，眼動資料也顯示讀者從文字與圖片所形成的空間表徵有所不同，且空間表徵會影響讀者在後續動態訊息文本的閱讀歷程：文字組讀動態訊息文本的總凝視時間和重讀時間都較圖片組短，特別是在動態訊息文本中概念較不複雜的句子（描述儲水槽的出水運作歷程）和儲水槽部件的目標詞上。綜上，本研究確認文字和圖片於傳達空間結構訊息的功能有所不同，亦說明讀者在閱讀過程中形成空間表徵與動態表徵的互動關係，進而建構文章內容的心智模式。

This study investigated how words and pictures convey spatial and kinematic representations in a reading condition and the relations between these two representations. Participants read a text (text group) describing the spatial configuration of a flushing cistern or viewed a diagram (diagram group) and completed a spatial test. Then, the two groups read the same text describing the kinematic machine operation and revised their first test answers. The participants’eye movements were recorded while reading the kinematic information text. According to the results, the text group’s drawings showed the component relations more precisely, indicating that words helped readers capture detailed continuous relations, but the diagram group’s drawings better approximated the original picture, indicating the advantage of depicting overall relations. Analysis of the drawing tests indicated that kinematic representations help readers revise inner spatial representations mainly of continuous relations of components rather than part-whole relations. Moreover, eye movement data indicated that the groups differed in both global and local eye movements, and that different internal spatial representations of the spatial and diagram groups influenced the following kinematic information text reading. The text group had significantly shorter total fixation duration and second-pass reading time on the kinematic text, fewer complicated concept sentences (i.e., outlet processes), and fewer target words for components than the diagram group. In summary, this study confirmed that pictures and words served fundamentally different functions in conveying spatial configurations and clarified the interaction of spatial and kinematic representations from reading to mental model construction.