中文词切分的依据及眼动控制模型的解释

The Basis of Chinese Word Segmentation and Explanation of Eye Movement Control Model

回顾字母文字研究发现,词间空格是有效的词切分线索。而中文没有词间空格,那么是什么线索起着字母文字中空格一样的作用呢?总结以往研究发现,汉字的统计属性、词汇可能性、词素位置、词素熟悉性等因素可能是中文阅读中的词切分线索。字母文字词切分依据物理线索,中文词切分可能依据"心理线索"。未来研究应该解决中文词切分的核心线索以及线索替代的可能性,并构建中文眼动控制模型来解释中文阅读过程及词切分问题。

In the process of reading, word recognition is the key for readers to acquire effective information. In the process of word recognition, we must first identify the language symbols in the text, and then combine these symbols into different levels of linguistic units, such as words, phrases and sentences, and to determine the relationship between the various units. This process is called word segmentation. In the writing system of most alphabetic languages, such as English, German and Spanish, there are spaces between words, and spaces act as a clear physical clue to mark the word boundary. Moreover, spaces between words provide readers with clear visual information so that the readers complete the word segmentation in the visual perception stage, and the probability of word segmentation errors is very low. Reviewing the alphabetic languages studies, we found that the space between words is valid for word segmentation. Space can help readers to determine the boundaries between words, and it can also promote word recognition and improve word processing speed. Nevertheless, Chinese is a kind of language based on characters. Different from any alphabetic language, there is no clear visual mark between words in Chinese text, and the spacing between words is the same. Each Chinese character can be either a single word or as part of a word or phrase. Depending on the context of the sentence, a Chinese character can be either an independent word or a part of a word. There are many ambiguous words in Chinese, which depends on the context to determine its meaning, and then determine whether it is an independent word or a part of the vocabulary, which makes the Chinese word segmentation more complicated. However, these do not affect the normal Chinese reading, indicating that Chinese reading must have its own special word segmentation basis, which plays the same role as the spaces in alphabetical languages. Because there is no space between the Chinese characters, it should be discussed what can be the same as the space in English to help Chinese readers to segment words. In the previous studies, it has been found that the statistical attributes of Chinese characters, vocabulary possibilities, morpheme position, morpheme familiarity and other factors may be the mark of word segmentation in Chinese reading. A latest research suggests that in the reading process, the text is read in a strictly continuous （perhaps object-based） way. As we know, the word segmentation of alphabetic languages is according to physical clues, such as the space between words, but the word segmentation in Chinese may be based on psychological clues. The process of word segmentation is also related to other factors, such as individual differences and preview effect. This article holds that the future research should resolve the core clues of Chinese word segmentation and the possibility of clue substitution, and construct Chinese eye movement control model to explain the Chinese reading process and word segmentation problem.