Study On CFL Learners’ Word Segmentation Mechanism In Chinese Reading ——Evidence From Eye Movements

Word,as a fundamental unit of sentence reading,plays an indispensable role in segmenting the sentence into meaningful units and integrating meaning effectively in a later phase of reading process.The awareness of word segmentation in Chinese is developed naturally in language acquisition for natives,but for Chinese as a foreign language（CFL）learners,it needs training.Alphabetic languages like English,German,Spanish are categorized as“spaced languages”since interword spaces are explicitly provided in their writing systems,facilitating the word boundary identification processes.However,languages like Chinese and Japanese,categorized as“non-spaced languages”,have no such explicit cues to mark word boundaries,resulting in more cognitive efforts that would be made to second language learners whose first language belongs to“spaced languages”.With the development of International Chinese Language Education（ICLE）worldwide,more and more foreigners learn Chinese as their second language.Today’s ICLE should be instructed by more scientific and powerful theories so as to be aligned with the learners’learning goals effectively.Evidence from empirical studies is able to verify current theories as well as to provide primary data to propose a new theory,which is the main objective of this study.In terms of reading process,it has been widely proved that introducing interword spaces to CFL learners would significantly facilitate their reading speed.However,is it necessary for us to introduce interword spaces to CFL learners from Chinese-character cultural circle countries（e.g.,Japan,North Korea,South Korea,Vietnam）,or to those from some southeast Asian countries（Singapore,Malaysia,Indonesia）where Chinese language is widely used?Until now,this question is not clearly answered.Therefore,this research tries to investigate the reading patterns of CFL learners,including CFL learners from non-Chinese-character cultural circle countries and those from Chinese-character cultural circle countries,and to compare with Chinese natives’reading patterns by using an eye-tracker to collect participants’eye movements,in order to find out the word segmentation mechanisms of CFL learners.The materials used as stimuli during the reading task are presented in three forms:the normal unspaced text,the word-spaced text,and the nonword-spaced text,designed to examine the form（s）that will facilitate or disrupt their reading processes.In order to achieve the goal,we propose the following three research questions:1)To CFL learners（including both groups from non-Chinese-character cultural circle countries and Chinese-character cultural circle countries）,which presenting form would facilitate or hinder their Chinese reading?2)Compared with CFL learners from non-Chinese-character cultural circle countries and those Chinese-character cultural circle countries,are there any differences between these two groups of CFL learners in their reading patterns for segmenting Chinese words?If yes,then why?and 3)Are there any differences in CFL learners’reading patterns compared with Chinese natives?If so,in what way（s）?If not,in what way（s）?The eye-tracking experiment of this research is a two-factor 3×3 mixed experimental design.Presenting condition（F₁）is a within-subject variable with 3 levels（the normal unspaced condition,the word-spaced condition,the nonword-spaced condition）,and participant group（F₂）is a between-subject variable with 3 categories（NC[CFL learners from non-Chinese-character Cultural Circle countries];CC[CFL learners from Chinese-character cultural circle countries];and CN[Chinese natives]）.The dependent variable is the individual participant’s eye-movement index.Regarding the indices being analyzed in this research,6 eye-movement indices are adopted:1)the total sentence reading time,2)the mean fixation duration,3)the mean saccade length,4)the total number of fixations,5)the total number of saccades,and 6)the number of regressive saccades.The general results answered the above research questions straightforwardly.Firstly,the word-spaced text facilitated reading for the NC participants substantially,while no significant effects were found for the CC participants.However,the nonword-spaced text impeded reading for all CFL learners.Secondly,for learners from non-Chinese-character cultural circle countries,their highest reading rate was found when they were reading the word-spaced text,while for those from Chinese-character cultural circle countries,they performed best under the condition of normal unspaced text.Regardless of the difference between cultural backgrounds,the reading rate of the two groups of CFL learners were both hindered substantially in reading the nonword-spaced text.Moreover,it was indicated from the data that 1)the total sentence reading time,the mean saccade length,and the total number of saccades were significantly different between the three groups,and the findings of interaction effects of independent variables indicated that both presenting conditions and participant groups influenced the participants’eye-movement patterns;2)the mean fixation duration and the total number of fixations also indicated pairwise differences between three groups,but without significant interaction effects;3)the number of regressive saccades produced interesting comparisons between groups,displaying that NC group differed significantly from either CC or CN group,while CC and CN groups did not show much discrepancy,moreover,no significant interaction effect was found to regressions.These findings answered the third research question that although the learners from non-Chinese-character cultural circle countries demonstrated a significantly different pattern of eye movements from the Chinese natives,the patterns of learners from Chinese-character cultural circle countries were similar to Chinese natives.The CFL learners’HSK（Hanyu Shuiping Kaoshi,a Chinese Proficiency Test）levels and their years of Chinese learning were also discussed with application of LME（Linear-mixed effects model）analyses,under such conditions that either their HSK levels or years of Chinese learning was included as a random effect in the LME model.The statistical results showed that neither of the two factors had significance on their reading patterns.Besides,a measurement for self-perceived sentence difficulty indicated that both CC and CN participants reckoned the sentences easy enough to understand,while NC participants assessed the sentences more difficult to understand,indicating that the normal unspaced text produced more perceptual burdens to NC participants.The general results have further proved that Chinese words do have psychological reality.Furthermore,the findings do provide pedagogical implications to CFL reading teaching methods in the future.Although the two groups of CFL learners’reading patterns were significantly different from each other,we need to treat the two groups of CFL learners with different teaching methods.For learners from non-Chinese-character cultural circle countries,esp.for the beginners,a word-spaced text would be more beneficial to successful comprehension than a normal unspaced text.While for the learners from Chinese-character cultural circle countries,who have already perceived Chinese characters before they began to learn Chinese,they would be faster in getting used to the correct Chinese words segmentation procedures,hence the normal unspaced text could be provided sooner in school teaching for this group of learners.