Conceptual Burstiness in Sociolinguistic Profiling of Radical-Criminal Communications: Corpus Method-Assisted Meaning Extraction for Investigative Leads

Abstract Positioned within linguistic profiling research as in Shuy’s work (The language of murder cases: intentionality, predisposition, and voluntariness. Oxford University Press, Oxford, 2014), this study introduces the concept of ‘conceptual burstiness’ to computer-assisted sociolinguistic profiling, demonstrating how to uncover investigative leads within terrorist threatening communications. Leveraging corpus analysis methods and semiotic clues, the study chiefly utilises AntConc (a software tool developed by Anthony, AntConc (Version 3.5.8) Computer software. Waseda University, Tokyo, Japan, 2019) to scrutinise word frequency and concordance lines, revealing distinct lexical preferences and their semantic proximity as recurrent thematic elements colouring the semantics of criminal texts. As a case study, the article examines 20 public statements attributed to notorious figures: the far-rightist Brenton Tarrant and jihadists Osama bin Laden, Abubakar Shekau, and Abubaker al-Baghdadi. Analysis conducted revealed repeated lemmas making up around 21.65% of the entire texts and encoding the semantic nature of the terrorist texts, giving rise to propelled conceptual burstiness categories as: (i) a function of the discursive purpose to which different lemmas converging within the same semantic fields are used in the forensic texts (e.g. introducing a radical concept and persuading the audience); and (ii) a reflection of the regularities in symbolic capital and repertoire of concepts to which the authors adhere. These discernible patterns reflect each author’s adherence to religious or ethnonationalist ideologies, patterns of violent-pathway language, agendas of control and dominance, and attribution practices that signal ideological stances and characteristic rhetorical devices. Integrating insights into the threat assessment framework TRAP-18 (Meloy & Gill, J Threat Assess Manag 3(1):37, 2016), the study offers a ‘post-diction’ (i.e. retrospective) lens on traditional risk assessment methods. The analysis of conceptual choices reveals patterns aligning with TRAP-18 categories of proximal warning behaviour and distal characteristics of terrorist communications, enhancing security and law enforcement professionals’ predictive ability and identifying pathways to radicalisation, fixation, identification with violent individuals and groups, desperation, or ‘last resort behaviour.’ The conceptual burstiness method promises insights for counter-terrorism efforts, informing accelerated investigatory and intervention protocols and objectionable content moderation online.