ABSTRACT
Data-driven learning (DDL), introduced in 1990 by Tim Johns, has come in multiple guises ranging from serendipitous corpus exploration to focus on pre-selected linguistic aspects. In the present chapter, I will show that despite the variety of approaches used, DDL is still mainly conceptualised as the analysis of written concordance lines and has not taken advantage of the numerous advances in digital technology over the last 30 years (e.g. multimodality, mobility, etc.).
After framing DDL in the light of some current theoretical frameworks of language pedagogy, I will discuss the norms, types of corpora, users, and technologies that are suitable for DDL (low tech, high tech, and wild tech options). I will show, with the help of concrete examples, that DDL can be revamped to make full use of the affordances of new technologies. Exploiting such affordances will hopefully also boost the use of DDL by and for many more learner profiles, such as younger learners or less favoured populations, thereby promoting more equitable multilingualism.
