视觉信息处理:阅读中文和英文时眼动模式的对比研究

文字阅读是一个复杂的视觉模式识别过程,虽然客观测量阅读过程中的脑的活动比较困难,但用记录分析阅读中由脑部控制的眼球活动来研究脑的信息处理特点却是可行的.本文采用两种不同结构的文字——以空间图象为基础的方块汉字和以发音序列为基础的横行英文——作为输入,用红外眼动仪测量阅读时的眼动曲线,再由计算机分析其异同.结果表明:1)阅读注视时程对中文及英文分别为0.260及0.265秒;2)阅读辨认跨度对中文及英文分别为1.7及1.8信息单位.由两者的相似性提示我们:整段理解阅读信息处理能力决定于神经系统高级阅读中枢的解码速率.

VISUAL INFORMATION PROCESSING: EYE MOVEMENTS DURING READING CHINESE AND ENGLISH

Reading is a complex pattern recognition process. From the viewpoint of information theory, reading can be considered as a decoding process. The visual language patterns are transformed by retina to neural signals, then decoded by the central nervous system, meanwhile the central nervous system also presents feedback control to the eye movement for sampling the information efficiently. It is difficult to measure the reading information processing in the brain directly, but it is possible to study the information processing in central nervous system by recording and analysing the eye movements.Two very different language codes--printed Chinese and English--werechosen to test the performance of reading information channel. The Chinese characters, with the boxlike shape are better adapted to exploit the disc fovea than the linear shape English words; the pictorial or phonetic-signif ic combination Chinese differs greatly from the linear alphabetical English. The experi-ment results show: 1)The average duration of fixation pauses during reading Chinese and English texts are very similar, i.e. 0.260 and 0.265 seconds respectively. It means that sampling and processing cycles of reading these two languages are very close. 2) The span of recognition for reading Chinese is 1.7 equivalent words, while the span for English is 1.8 words. The reading rate are 380 and 390 equivalent words for Chinese and English.The similarities between the eye movement pattern in reading Chinese and English indicate that information processing capabilities of the brain are chief limiting factor.