The Processing Of Subjective Words In Chinese Reading

In alphabetic language, there are spaces between words and readers can use thisboundary information to read, so word is considered as the basic information unit. However, no apparent visual cue can be used to segment text into words in Chinese reading. What’s more, Chinese readers learn less about word and even don’t know "what is a word". Then there is an interesting question:what is the basic information unit in Chinese reading? Lots of studies have explored this question, but no consistent conclusion on it. This might be caused by the co-existence of all kinds of units. Yan, Zhang, Sun and Bai (2013) provided a new conception:subjective word and suggested that subjective word is the basic information unit. Given the meaning of subjective word, it is more flexible and variable and takes the readers into account. However, how the subjective words being processed in the normal reading is still unclear.In the current study, three studies were conducted to explore the processing of subjective word and look for the relationship between critical properties of compositions with the whole subjective word.Study 1. In this study, we investigated the influences of subjective word processing on Chinese reading. A highlighting method was adopted to mark the boundaries of character, word and subjective word. At the same time, the individual difference was included. There were two stages in the present study. First stage (in the beginning of the term), participants were asked to segment sentences into word according to their understanding, so we can get the subjective words of theirs. Second stage, some of them were asked to complete the reading task and eye tracking system was used to record their eye movements during reading. Thesentences were marked with four kinds of boundaries:character, word, subjective word and no boundary. Under the subjective word condition, the boundaries were marked according to the segmentation in the first stage. The results showed:The mean saccade length and forward saccade length under subjective word condition were longer than those under character condition, but no significant difference on reading time among four conditions. we considered that the reason why no improvement on reading speed was there were some subjective words can’t be processed as a whole, so we selected some target 4-character combinations which were segmented as one word by most of participants. The data on target combinations showed a robust advantage for subjective word condition over character condition on early measures.According to the result of study 1, we considered that there were different kinds of subjective words and the processing would also be different. So we divided the subjective words into three categories according to the proportion of being considered as one word:word (they are words according to the CCLWSSIP and above 80% participants also considered them as one word), definite subjective words (they are phrases according to the CCLWSSIP, but above 80% participants considered them as one word), ambiguous subjective words (they are phrases according to the CCLWSSIP, and 40%-70% participant consider them as one word).Study 2. The main object is to investigate whether there are advantages when process definite and ambiguous subjective words? 3 experiments (2-4) were conducted to explore this question. In experiment 2, we explored the foveal processing of definite subjective words. Pairs of 4-character strings (definite subjective words, non-subjective words) were embedded in the same sentence frames, participants’ eye movements were record during reading sentences. We found that: there were significant differences between the processing of definite subjective words versus non-subjective words, the definite subjective words showed fairly strong advantage on fixation times. What’s more, the processing of subjective words’ components even faster. In experiment 3, we studied the processing of ambiguous subjective words. Participants were firstly asked to read sentences with 2-word ambiguous subjective words, their eye movements were monitored, and then complete a word segment task. These participants were divided into two groups according to their segmentation on subjective words. We compared the reading time on ambiguous subjective words of two groups, but no significant difference was found. These results suggested that the ambiguous subjective words might not be processed as a whole. In order to testify thisassumption, experiment 4 was conducted. We selected three kind 4-character strings (definite subjective words, ambiguous subjective words and non-subjective words), and embedded them in the same sentence frames, participants’eye movements were record during reading sentences. We found that:compared with the non-subjective words and ambiguous subjective words, the fixation times were shorter, fixation counts were less. However, there was no significant difference between the processing of non-subjective words and ambiguous subjective words. These results showed that the ambiguous words has no processing advantage, and can’t be processed as a whole.Study 3. The results of study 2 suggested that the definite subjective words should be processed as a whole, so the experiment 5 was conducted to study how the constitutes were processed, parallel or sequential? A boundary paradigm was used to compare the parafoveal processing of definite subjective words and the non-subjective words. We found a significant parafoveal and foveal effect on the first constituent of definite subjective words and preview benefit on the second constituents of two kinds of strings. These results meant that the components of definite subjective words were processed in parallel, so they can be identified by the look-direct up entry.To sum up, there are different kinds of subjective words and the processing are variable:the definite ones can be processed as a single unit, while the ambiguous ones still are processed as two parts as non-subjective words. As a result, the definite subjective words can be the basic information unit.