| 英文摘要 |
Geometric texts carry rich spatial information and arguments. This study examined the eye movement of adults reading different types of geometric text to explore Duval’s model of visualization and reasoning. A total of 50 participants were requested to read two geometry proof items and nine geometric description items. Eye movements during proof reading were categorized into “given,” “prove,” and “proof examples.” Three types of geometric text comprised different proportions of reasoning components. Answers to the question of to what extent reasoning elements are involved in the three types of geometric text were ranked in the following order: “geometric description,” “given and prove,” and “proof example.” First, we found that the total fixation duration ratios of the graph for these three types of geometric text were 36%, 41% and 51%, respectively. The ratio of proof examples was significantly higher than that of the other two types of geometric text, indicating that the extent to which readers rely on figures when reading geometric texts is influenced by differences between the three types of geometric text. Second, according to the total fixation time, the process of reading comprehension for a geometric text can be divided into two phases: initial comprehension and rereading. As indicated from the reading time ratios and average reading time for areas of interest (AOIs), we observed that initial comprehension could be related to Duval’s visualization, and that rereading could be related to Duval’s reasoning. Finally, based on an analysis of the Markov transition matrix, geometric symbols—particularly keynote geometric symbols—were regarded as clues that guide the reader from the text to the figures; however, word description, geometric terms, and geometric symbols embedded in the background were proven to direct readers to move onto subsequent AOIs in the text, but rarely onto figures. |