As the numbers of those with an interest in terrorism increase – could we say there is a boom in the field? – the number of novel approaches also increases. We have seen studies done on social media postings, attempts at understanding the psychology of extremism and myriad tries at detecting inputs (i.e. aspects of a person’s life before they embrace violence). Some of these have contributed to the picture we are trying to paint and all have asked some good questions, even if they have not brought usable solutions.
A new idea in Sweden has caught my eye as it combines terrorism and Big Data. Many people see the collection and exploitation of Big Data as an incredible opportunity to gain insights into all kinds of human behaviour (I reserve judgment for now). Researchers at the Swedish Defence Research Agency are trying to use how people write about themselves online to see whether they can detect signs of impending violence. Their theory states that there may be certain “writing prints” that betray the common origin of several created online profiles and there may be clues in the language used that can help security and law enforcement in their efforts to identify and prevent terrorists from acting (read a short article on this project here).
There are a few things about this project that I like, as well as a few cautionary notes. First and foremost I love the humility of the researchers. They acknowledge that their work is perfectly theoretical at this point and that they have no idea how practitioners would operationalise it (as a former practitioner thank you for not claiming to have the answer to everything. I also think Scandinavians and Canadians may share a humbleness gene!). They also are all too aware of privacy issues and want to carry out their study in public to avoid the fear of being labelled Big Brother. Well done, even if the Big Brother thing is a little overdone in my opinion.
Now for the cautionary notes – don’t worry they are minor. I would be worried about the incidence of false positives. It was my experience that many more people talk about terrorism than actually do anything about it. How would this research tease the two apart? Would “talkers” be confused with “walkers”? Secondly, I am not sure there are linguistic clues that can be detected (full disclosure: I taught linguistics, including a course in psycholinguistics, for over a decade). What would a person have to say, and how would they have to say it, to be of concern? This research question as developed by the Swedes has really piqued my interest. Every person has their own idiosyncratic way of using language but we also have multiple “languages”, each appropriate to a particular social environment (theses are called “registers”). Do people online always use language in the same way? I doubt it (I know that my FaceBook postings are very different than my Tweets for example). Do extremists betray their true intent without knowing it? Great question.
I am looking very forward to the results of this research. If nothing else, it will provide one more small insight into the complex problem that is terrorism.
