The Mechanism Of Word Segmentation During Chinese Reading:An Eye Movement Study

Word is generally considered to be the smallest basic processing unit, which could be used independently. Unlike alphabetic writing systems, Chinese text is written with continuous characters, without spaces between words. That raises a basis but important question:how Chinese readers segment text into words during reading Chinese text? The exploration of this issue is not only the key point of the cognitive mechanism of Chinese reading, but it also aids to develop the technique of word segmentation in the area of linguistics. Furthermore, the solution of this issue will contribute directly to the development in the field of computer translation, automatic speech recognition, and human-machine dialogue.On the basis of the previous studies, the present study aimed to explore all the possible word segmentation cues during Chinese reading, especially we explored how these cues operate together on the process of word segmentation, in order to make clear with the basic processing unit of Chinese, and provide data and theoretical basis for the construction of the model in Chinese reading.Two studies including three experiments were carried out. Study1explored what were the factors could influence the effects of word spacing, as a low-level visual word segmentation cue when reading and learning new vocabularies within sentential contexts. Based on the findings of Study1, the next study further investigated how two different types of word segmentation cues (word spacing as a low-level visual segmentation cue and character’s positional frequency as a high-level linguistic word segmentation cue) operated on the process of word segmentation and word identification when learning new vocabularies during Chinese contexts. The detailed contents and main findings were shown as follows.There were two experiments in Study1. In experiment1, we investigated whether the interword spacing effects was modulated by the ease of the new vocabularies. We constructed two types of two-character pseudowords as experimental target new vocabularies in accordance with the familiarity of the constituent characters composed of the new vocabularies. All the target words were put into several sentences, each of which provided a context to comprehend the novel words. In experiment2, we further investigated whether the familiarity of constituents with different positions in words modulated the facilitation effects of word spacing during the learning of new Chinese vocabularies.In experiment3, three types of pseudowords were constructed based on each character’s positional probability, providing congruent, incongruent, and no positional information respectively. Here, we aimed to address the two following questions:(a) whether visual and linguistic cues were used concurrently, or if one type of cue dominated the segmentation process; and (b) if visual and linguistic cues were found to be used concurrently when reading, then whether those cues had an additive or an interactive relationship.All the three experiments in the present study adopted the same experimental paradigm, and the undergraduates were selected to take part in all the experiments. To summarize the findings of the three studies, the conclusion was made as follows.(1) There was a typical word-spacing effect on the learning of new Chinese vocabularies during sentential contexts. The insertion of spaces between words facilitated the process on the formation of novel lexical representation. Once the novel word representation had been formed, the facilitation of word spacing would disappear.(2) The effect of word-spacing on the acquisition of new vocabularies was modulated by the ease of novel word itself. Compared with the novel words comprised of two familiar constituent characters, word spacing played less facilitation than those novel words comprised of two unfamiliar constituent characters.(3) The familiarity of constituents with different within-word positions was another factor to modulate the degree of word spacing effects when reading novel words during Chinese texts. Compared with the unfamiliarity of the constituent as being the word ending, word spacing was of more importance than when the unfamiliarity of the constituent as being the word beginning. (4) Chinese readers adopted the character’s positional frequency as one type of linguistic word demarcation cue to segment and identify the word as well. This type of word segmentation cue operated on the formation of novel word representation. In addition, it still worked after the lexical representation had been formed.(5) Word spacing and character’s positional frequency, as two different types of word segmentation cues, operated independently, rather interactively. Compared with the word spacing, character’s positional frequency worked with a longer duration and a larger range.