What
Threatening messages are an increasing menace in modern society causing high levels of personal distress by aiming to instil fear in their recipients. This is an effect largely produced by language, but very little is known about the linguistic characteristics of threats. This project studies the language and genre of threats through a corpus of authentic threatening messages. Scientifically, the project promises new insights into 'illicit genres' (un-institutionalised and socially proscribed text types). Further, it studies how threats are recontextualised in the legal documents of indictments and verdicts to become lawfully prosecutable offenses.
Why
Threatening messages belong to the most common person-based crimes. A better understanding of the generic and linguistic features of threats will advance the frontiers of basic research and introduce new applications for forensic threat investigation and assessment. Studies of English-language threats, show that lay people's and intelligence agencies' conceptions of threatening language focus too much on salient-but in fact genre-atypical-features such as swear words while overlooking the more widespread use of downtoning features. Therefore, detailed studies of the use of a range of linguistic features are necessary for establishing the genre-typical patterns of threatening messages. With sufficient background info on authors, results may be used for profiling anonymous threateners.
How
By building a corpus of threatening messages, we can study linguistic patterns in the data that will feed into analyses of their rhetorical functions. The threatening messages will be digitalised and tagged automatically for word class and syntactic function by Natural Language Processing tools. Automatic codings will then be corrected and augmented manually, based on results from international studies.
The project group interacts with internationally leading experts throughout the project to adopt and adapt best-practice solutions to Danish data.
SSR
The most important applications of this research are:
1. the possible improvement of threat assessments through a better understanding of the language of realised and unrealised threats (that is, threats that new acted upon or not acted upon).
2. the possible augmentation of criminal investigations through a later-stage implementation of the corpus of threatening messages into a searchable database for Danish police. This may be used to compare threat cases across the country to locate multiple offenders that span more than one police district.