Most borrowable construction ever! A large-scale approach to contact-induced pragmatic change

Overall, our account complements anglicism research (Andersen 2014) in four ways. First, we foreground constructional borrowing instead of single-word borrowing. Second, in working with Twitter data, we break with the tradition of print media corpora. Third, we explore NLP based methods for large datasets sampled from big data collections in a field of research that has mainly relied on manual coding of small-scale datasets. Finally, we illustrate how matter and pattern replication can go hand in hand in contact-induced change.