On Designing Interlanguage Corpus of Indonesian Students with Error Analysis Annotation

Abstract

This paper proposes a model for the design of interlanguage corpus with error analysis annotation. The data is obtained from ICNALE (Ishikawa, 2013) corpus, a corpus of English learners in Asia. I focus the extraction on the Indonesian data. The sample for this model is obtained from English texts written by proficient Indonesian students. The data is processed by a Local Grammar (LG) based corpus processing software called UNITEX (Paumier, 2008). The experiment begins with error identification. Parts of the texts where errors take place are marked with error tags. The suggested forms to revise the errors are written in Local Grammar Graphs (LGGs) that are applicable to the corpus text. When these LGGs are applied in the corpus, they retrieve parts of the texts that are marked by the error tags and the suggested forms embedded in the LGGs. Users are provided with the options to display concordance in several ways. First, it may identify parts of the texts that are marked by error tags. Second, it may identify the error and display the suggested forms as well. Third, it allows the suggested forms to replace the error tags. Hence, the errors are hidden. This corpus may benefit teachers and linguists in terms of material development and L1 interference-analysis support system.