Reproducibility in corpus linguistics
What is it about?
In many fields, there is increasing concern about the credibility of scientific research. However, it is not entirely clear what it means for a scientific result to be deemed credible. This is particularly the case in research involving language research, especially those that combine quantitative as well as qualitative methods.This paper examines the different facets involved in what makes a result credible in the field of corpus linguistics, laying out the different considerations researchers should keep in mind.
Why is it important?
Questions of credibility affect all sciences, but the issue has remained unexplored in corpus linguistics, despite the fact that many of the concerns of the credibility movement can be found since its inception. While the paper doesn't address directly the credibility of large language models, its delineation of what it means for research in language to be credible is also certainly relevant to that discussion.