Most models of the compound words recognition are based on the effect of frequency.The word frequency effect reflects the representation of the whole compound word and thefrequency effect of morphemes reflects the representation of morpheme. According to thetime-course of frequency effect, some researchers have different understanding for therecognition of the compound words, and constructed different theoretical models. Manyresearchers believe that the effect of word frequency facilitates to the identification of thecompound words, but there are divarications regarding the interaction among the wordfrequency, the frequency of morphemes and the family size of morphemes. We also found thatthere are some differences on methodology in existing research. On the one hand, someresearches neglected the interaction among word frequency, frequency of morphemes and thefamily size of morphemes. On the other hand, some researches neglected the influence of theduration of stimulates presented in experiments studying compound words. This study willexamine the interaction among word frequency, frequency of morphemes and the family sizeof morphemes, and discuss the recognition mechanism of Chinese compound words andfrequency effect.Experiment1, the purpose of which is to compare itself with the results of some previousresearches (Peng, et al,1999; Tian, et al,2009), simulated previous experiments’ conditions,using the same experiment task but different presenting duration of stimulants(2000ms,1500ms,1000ms,500ms and300ms). The results show a small facilitating effect of familysize and inhibitory effects of the morpheme frequency. The word frequency induces the effectof morphemes family size and the effect of words frequency regulates the processing of themorpheme representation. The relationship between the three factors is complex and dynamic.We also found that the family size of morphemes occurred at the earlier stage of theprocessing of compound words. The presenting duration of stimulants is an important factorinfluencing the effect of frequency and a cause for the difference in the results of the previousresearch. Experiment2probes into the relationship of the frequency among the frequency ofmorphemes, the word frequency and the semantic transparency. Like the former experiments,experiment2is also carried out using the lexical decision task. The results indicate that thereis an interaction effect among the frequency of morphemes, the word frequency and thesemantic transparency. The results suggest that semantic transparency and word frequencydefermine the direction of the morphemes frequency effect. In experiment2a-2c, the resultsalso validate the influence of the presenting duration of stimulants on the effect of frequencyfactors. The results of experiment2d indicate that the priming effect of word semanticrelevancy appears for43ms in both high word frequency compounds and low word frequencycompounds. It is indicated that the representation of whole word can also appear in the earlystages of the compound word recognition. The results of experiment2d support the view thatthe semantic transparency and the word frequency determine the direction of the morphemesfrequency effect in the early stage of compound words recognition.In experiment3, we used tracking technique with high time precision of eye movementto investigate the relationships among factors for the same experiment materials asexperiment1. The interaction between the frequency and the family size of morphemes ismarginal significant for the first fixation duretion. One of the important results is that theeffect of the family size of morphemes appears very early. Throughout the experiment, we canmake clear the relationship among word frequency, morphemes frequency and the size ofmorpheme family.Using the tracking technology of eye movement, the experiment4checked therelationship among the frequency of morphemes, the word frequency and semantictransparency. Like the former experiments, experiment4is also executed using the lexicaldecision task. The results indicated that (1) the frequency of morphemes reflecting morphemerepresentation influences the duration of first fixation,(2) semantic transparency and wordfrequency influence the duration of second fixation.Experiment5uses the tracking technology of eye movement to check the relationship among semantic transparency, word frequency and the frequency of morphemes of theChinese compound words in sentences. The results indicate that (1) semantic transparencyinfluences the duration of first fixation in the first run,(2) the frequency of morphemes,semantic transparency and the word frequency influence the duration of the second fixation.The results in this paper show that the effect of the family size of morpheme and theeffect of the word frequency appear earlier. Though the earlier process is very short, theinformation of the compound word is perceived very quickly and effectively. This paper alsofound that the representation of the whole compound word is in a dominant position duringthe lexical access. The important factors for compound word recognition are founded justduring the first scan of our eyes. Before precise processing of compound words, therecognition system must make a rough processing. The readers cast their first look at thecentral area of the compound word, basing on a lot of reading practice and vocabularylearning knowledge. In the first scan, the recognition system would assess the information andidentify the most important information. The recognition system will follow the clues andregulate the lexical information representation activation level, and collect the availableinformation to promote the representation activation level of whole word to assist theconfirmation of the compound words and complete the recognition of the words. |