Toward Online Linguistic Surveillance of Threatening Messages
Brian H. Spitzberg, Jean Mark Gawron · The journal of digital forensics, security and law · 2016
Threats are communicative acts, but it is not always obvious what they communicate or when they communicate imminent credible and serious risk. This paper proposes a research- and theory-based set of over 20 potential linguistic risk indicators that may discriminate credible from non-credible threats within online threat message corpora. Two prongs are proposed: (1) Using expert and layperson ratings to validate subjective scales in relation to annotated known risk messages, and (2) Using the resulting annotated corpora for automated machine learning with computational linguistic analyses to classify non-threats, false threats, and credible threats. Rating scales are proposed, existing threat corpora are identified, and some prospective computational linguistic procedures are identified. Implications for ongoing threat surveillance and its applications are explored.