Exploring a Corpus-Based Approach to Assessing Interpreting Quality

Abstract

Interpreting quality assessment (IQA) is a challenging task, due to possible bias of human raters and transient nature of interpreting. To improve the consistency of IQA, researchers can draw on corpus linguistics to analyze machine-readable interpreting corpora. A corpus-based approach can potentially provide a more consistent way to evaluate interpreting quality. In this chapter, the author describes such an approach to evaluating the quality of students’ interpreting performance. A total of 64 English-to-Chinese interpreting renditions were sourced from the Parallel Corpus of Chinese EFL Learners (PACCEL) and then profiled based on linguistic features extracted from the corpus. These linguistic features were grouped into three quality parameters: information accuracy, output fluency, and audience acceptability. Principal component analysis and decision tree analysis were conducted on the multi-dimensional linguistic data to identify the appropriateness of proposed assessment indicators, and to verify assessment accuracy. Finally, assessment results were visualized via kernel principal component analysis. The results indicate that the proposed approach was capable of distinguishing students’ interpretations of different qualities. The study also shows that corpus linguistics has the potential to contribute to the development of IQA research.