“This is an extortion note”: a corpus-driven genre analysis of commercial extortion letters

The paper presents a corpus-driven analysis of a series of 39 commercial extortion letters and emails from historic cases in the UK (2008-19), called the Excrow corpus (Extortion CoRpus Of Writings).
Using Swales’ (1990) Move Analysis, we explore whether conventional discourse structures of a genre (i.e., moves) can be identified in extortion letters. We then analyse the identified moves with corpus linguistics and clustering algorithms.
The paper presents two major innovations. Firstly, we develop a reliable and replicable bottom-up method for corpus-based Move Analysis, taking clauses as basic units of analysis and employing inter-rater reliability tests. In this way, a set of 11 key moves in the data. Secondly, we use the set of moves to conduct quantitative corpus-driven (n-grams and sequence analysis) and computational analyses (using clustering algorithms). Results indicate a high degree of variability in move sequences, and no obvious recurring move patterns; however, we were able to cluster the letters in coherent and stable groups based on move prevalence.