Automatic profiling of learner texts

Abstract
In this chapter Crystal's notion of 'profiling', that is identification of the most salient features in a particular person or register, is applied to the field of interlanguage studies. Automatic profiling can help researchers form a quick picture of the interlanguage of a given learner population and that it opens up interesting avenues for future research. The automatic profiling technique has highlighted the speech-like nature of learner writing. There is also a very significant overuse in the learner corpus of the first and second personal pronouns. In the French learner corpus, the indefinite article a is overused and the definite article the underused. The overall underuse of nouns that characterizes French learner argumentative writing is thus clearly a further sign of a tendency towards oral style. Automatic profiling applied to a wide range of learner corpora has the potential to help us answer the questions and thereby contribute to a better understanding of learner grammar and lexis.