To Err is Human: An error analysis approach to Turkish as an L2

Abstract

In this study we investigate which aspects of Turkish pose particular challenges for English-speaking learners. The data are from a large pool of university LCTL learners who responded to five pre-recorded speaking prompts. They audio-recorded their speech twice a semester for up to three consecutive semesters using a virtual interview assessment tool called VOICES. Each interview lasted about 10 minutes. We identified and analyzed the various types and frequencies of grammatical and lexical errors that emerged in the Turkish speech from twelve English-speaking learners of Turkish. The results indicate that English-speaking learners of Turkish had problems with the Turkish case-marking system, subject-verb agreement, singularity/plurality, near synonyms and lexical shifts. Consistent with previous L2 Turkish studies, we found that some of the errors can be attributed to the fact that Turkish is an agglutinative language while English is isolating. We discuss how analyses of this type can improve LCTL instruction.

Keywords

• error analysis
• Turkish as a second language
• virtual interview