Corpus-based analysis of second language writing errors and their pedagogical implications

Second language writing is not only a major problem for language learners but also a long-term focus of attention in the field of applied linguistics, and the error analysis of second language writing is of great significance for the development of language education at different stages. Multimodal discourse analysis is used to construct a corpus and propose a research methodology for second language teaching in this study. The errors in writing for second language teaching are investigated in detail based on the theory of error analysis in five steps: collecting samples, identifying errors, describing errors, interpreting errors, and evaluating errors. Through the investigation and analysis, it was found that among the four dimensions of functional adequacy, the task requirement dimension had the highest average score (5.71) and the most concentrated data, and the average scores of the comprehensibility (4.55) and coherence and articulation (4.35) dimensions decreased in order. Among the linguistic feature indicators, lexical complexity, lexical diversity, syntactic complexity and writing scores all have significant positive correlations, with correlation coefficients in the range of 0.26 to 0.35 and p-values less than 0.01. Among them, lexical diversity and writing scores have reached a moderate correlation coefficient of 0.35, which indicates that the more lexical varieties there are, the more low-frequency words are found, the more words are included in an average of each T unit, the more words, the higher the overall rating of the essay. The amount of writing errors in the student essay samples was all negatively correlated with the student’s essay scores, and the p’s were all less than 0.05, which is a very high correlation; the higher the number of writing errors, the lower the essay scores. This dissertation makes recommendations for teaching second-language writing and provides a reference for research on teaching writing in a second language based on the analysis results.