1. Introduction

This article introduces the structure of filler vacancy dependency and its importance in psycholinguistics, and explores two theoretical hypotheses about the establishment of a dependency relationship between filler and vacancy: the positive filler hypothesis and the direct association hypothesis. Both hypotheses suggest that filler language and verbs establish a dependency relationship before the emergence of arguments, resulting in a vacancy filling effect when arguments appear. Distinguishing between these two hypotheses is of great significance for understanding language processing mechanisms and the differences between second language and first language processing mechanisms.

2. Review of relevant research

This article summarizes two main hypotheses in the study of filler vacancy dependency structure processing: the positive filler hypothesis and the direct association hypothesis. The positive filler hypothesis advocates that language processors predict vacancy positions based on the syntactic features of filler language, while the direct association hypothesis suggests that parsing is based on known information, and the filler vacancy dependency relationship is established after obtaining semantic structural information of verbs. Research on first language processing shows that native speakers mainly rely on syntactic information, while there is controversy over whether second language learners adopt the same strategy. The shallow structure hypothesis suggests that second language learners primarily rely on lexical semantic information for shallow processing, but other studies have shown that second language learners can also utilize syntactic and semantic information. The semantic rationality mismatch effect and object vacancy filling effect used in the study can explain the language processor's prediction of vacancies, but they are not sufficient to prove the strategies adopted. The verb transitive mismatch effect provides a new perspective for distinguishing between the two hypotheses. Staub's (2007) study supports the direct association hypothesis, while Omaki et al.'s (2015) study supports the positive filler hypothesis. Minemi U0026 Yano's (2021) study on second language processing did not find any verb transitive mismatch effects, supporting the direct association hypothesis. However, there are shortcomings such as contradictory hypotheses, biased data analysis, and failure to consider working memory. The aim of this study is to explore the processing mechanism of Chinese English learners through the effect of verb transitive mismatch, and to examine the influence of English proficiency and working memory.

3. Research Design

A research design was introduced to explore the processing strategies of Chinese English learners towards the filler space dependency structure. The experiment adopts a four factor mixed design, including verb type, sentence type, English proficiency, and working memory level. The experimental materials were selected from Minemi u0026 Yano (2021), with a total of 32 key stimuli and 4 experimental sentences in each group, divided into 14 reading areas. The test sentence uses a parallel predicate structure with the same verb transitivity to test the influence of verb transitivity. The experimental and filler sentences are presented to the subjects through pseudo-random processing, using a moving window word by word masking paradigm. After reading each sentence, participants will take a true or false test to assess their comprehension of the sentence. The experiment was conducted on a desktop computer in the laboratory using E-prime 2.0 software, and it takes approximately 30 minutes to complete the entire experiment. The participants were 103 college students from a university in Nanjing. According to the scores of the Oxford Rapid Placement Test and Working Memory Capacity Test, they were divided into high and low English proficiency groups and high and low working memory groups. The data analysis methods include logistic regression mixed effects model and linear mixed effects model, which analyze the response time data of key regions to reveal the strategies of participants in constructing fill language vacancy dependency relationships.

4. Research results

Analyzed the accuracy of judgment and reading reaction time of the subjects under different testing conditions. The results showed that the average judgment accuracy of the subjects was 78.6%, and the judgment accuracy of the high English proficiency group was generally higher than that of the low English proficiency group, while the working memory level had no significant effect on the judgment accuracy. Logistic regression mixed effects analysis shows that English proficiency has a significant impact on judgment accuracy, while the interaction effect between working memory level, verb type, and sentence type is not significant. Simple effects analysis found that the accuracy of judgments under non relative clause conditions is generally higher than that under relative clause conditions. In the analysis of reading reaction time, erroneous data, extreme values, and invalid values were excluded, and outliers were processed. The analysis of key regions 8 and 9 shows that the reaction time under relative clause conditions is significantly longer than that under non relative clause conditions, while the analysis of region 10 indicates that when the verb is transitive, the reaction time under relative clause conditions is significantly longer than that under non relative clause conditions. However, when the verb is intransitive, the difference in reaction time between the two sentence types does not reach a significant level. Overall, sentence types have a significant impact on judgment accuracy and reading reaction time, while the effects of English proficiency and working memory level differ.

5. Discussion

Explored the processing strategies adopted by Chinese English learners when dealing with filler vacancy dependency structures. Research has found that when the key area 8 is a verb that is inferior to the object, the reaction time under relative clause conditions is significantly longer than under non relative clause conditions, supporting the positive filler hypothesis and indicating that learners adopt filler driven strategies. However, under transitive verb conditions, the reaction time of relative clauses is significantly longer in key regions 8 to 10 than in non relative clauses, consistent with the findings of Staub (2007) and Minemi u0026 Yano (2021). However, the authors believe that this cannot be directly inferred as a verb driven strategy. In online processing, the storage and integration of filler language require cognitive resources, so the longer reaction time under relative clause conditions is predictable. In addition, when the test sentence adopts the "V1 or V2" structure, it was found that the storage cost of filler language did not significantly affect the response in key region 10, which is inconsistent with the verb driven hypothesis but consistent with the filler language driven hypothesis.

6. Conclusion

Summarized that Chinese English learners are able to use the same proactive filler language strategies as native English speakers when dealing with filler vacancy dependency structures. The research results do not support the shallow structure hypothesis, indicating that second language learners can master the syntactic generation knowledge of English relative clauses through classroom teaching and internalize it as an effective online processing mechanism. This provides important insights for second language English teaching, namely that classroom teaching helps second language learners acquire explicit grammar knowledge and gradually transform it into implicit grammar mechanisms.