Investigating writing development, cross-linguistic influence and feedback practices through a longitudinal corpus of children’s school writing

Abstract

This article reports on our work with child writing from the TRAWL Corpus, a longitudinal and multilingual corpus of school writing. We give examples of our work on vocabulary and complexity development, which utilizes the longitudinal design of the corpus, and on feedback practices and student uptake facilitated by the fact that many of the texts in the corpus include teacher comments. These studies illustrate how corpus data can be used in case studies and qualitative studies and emphasize the need for fine-grained classifications in learner corpus research. The TRAWL Corpus includes texts written in the five languages most commonly taught in Norwegian schools. We explain how our work on syntactic complexity and feedback practices is currently being expanded by an exploration of similarities and differences between language subjects and of interactions between the languages of individual learners through the project MULTIWRITE. The article highlights the benefits of corpora of authentic school writing that reflect the realities of the educational context and therefore can provide findings that are directly relevant and useful for the practice field