What is data-driven learning?

Typically, course books present and practise grammar, vocabulary, language functions, pronunciation and ‘the four skills’: reading, writing, listening and speaking, packaged together in thematic units. Course-book writers therefore need access to a corpus of the chosen variety to ensure that their syllabus and language models accurately reflect usage in the corpus. If the trend towards corpus-informed materials continues, it is hoped that writers of materials for younger learners will be able to draw on corpora of young people’s language, in order to present age-appropriate models. As successful course books generally fit within an established body of knowledge in terms of language description, the course-book writer may be more inclined to the latter. Course books have generally borrowed little from the explicit use of corpus data seen in data-driven learning. In summary, there is bright future for corpus-informed course materials. A corpus of real conversations can provide excellent raw material for course-book presentations.