Design and construction of the Tracking Written Learner Language (TRAWL) Corpus: A longitudinal and multilingual young learner corpus

This article describes the design and construction of the Tracking Written Learner Language (TRAWL) Corpus. The corpus combines several features that are all rare for learner corpora: it is longitudinal, following individual pupils over several years; it has data from young learners from school years 5 to 13 (ages 10–18); it is multilingual, containing learners’ texts in several L3s (French, German and Spanish), L2 English and L1 Norwegian; and it includes teacher comments on a number of the texts. In addition, some of the texts exist in both a first and a second revised version, all tied to a rich set of meta-data. Not only does such a corpus offer new possibilities for research on language acquisition in general; it can also be used to provide valuable insights for teachers, teacher training and policymaking within the national context of Norway. In this article, we describe the design of the TRAWL Corpus and outline its uses and benefits for the research community. We also describe the compilation process in the hope that it may inspire and enable others to build similar corpora for their own national contexts.