Errors Versus Mistakes: A False Dichotomy?

Abstract

In this paper, the dichotomy between errors and mistakes in applied linguistics will be discussed, with a view to arriving at a clearer understanding of how to describe learner language from an empirical viewpoint. Errors are usually defined as systematic deviations from the rules of a target language. They may occur because a learner does not know a rule, such as Subject-Verb Agreement in English. Mistakes, on the other hand, are seen as unintentional, accidental slips resulting from simple laziness or forgetting. Distinguishing between errors and mistakes has always been fraught with problems. Using examples from a corpus of written learner English constructed in Malaysia, this paper argues that it is not empirically feasible, or even desirable, to maintain a dichotomy between errors and mistakes, especially from a corpus-based empirical perspective on learner language. It is argued that empirical analysis of learner language cannot practically access the knowledge of a learner to determine if an error or a mistake has been produced. Furthermore, many existing objective tests of ‘errorness’ or ‘mistakeness’ are not reliable. Instead, it is proposed that the phenomena currently known as ‘errors’ or ‘mistakes’ should be termed ‘interlanguage features’, a term which is relatively value-free.