Emails, SMS, letters and tweets reveal a lot about their authors to experts. This fact can help solve crimes.
Was the SMS truly written by the murder victim and was he or she still alive when it was sent? Or did the murderer send it after killing his victim? What might sound like the plot of a film is of great interest to linguists. Forensic linguistics strives to identify the authors of written texts or, at the very least, to narrow down the options.
Style features and characteristic mistakes
“We compare different properties of written texts, for example the choice of words, the complexity of the syntax, and the idiosyncrasies of punctuation. Another revealing aspect are characteristic mistakes that a person tends to make,” explains Steffen Hessler, PhD students at the SecHuman Graduate School. Together with Benedikt Bönninghoff, he pursues his PhD thesis in forensic linguistics at the Institute headed by Prof Dr Karin Pittner and has recently completed an internship at the Federal Criminal Police Office (Bundeskriminalamt, BKA). There, he studied genuine texts written by criminals and the relevant expert assessments. “Even though we used samples from old cases, reading the letters could be a harrowing experience,” he says.
Steffen Hessler has looked into the linguistic features that are used by BKA experts as reference when compiling an assessment. Benedikt Bönninghoff has integrated them into a neuronal network that is supposed to automatically identify them in a text. “Blackmail is the most common type of letter the BKA deals with, and today such letters often arrive by email,” relates Hessler. In order to match them with an author, it is necessary to consult a reference document that had been definitely penned by that person. “In one case, it was possible to match an old bomb threat, which was written on paper, with an author who has recently sent a bomb threat by email.”
Any attempts to disguise oneself are doomed to fail
Any attempts to disguise one’s style when writing a blackmail letter are doomed to fail Affecting deliberately sophisticated language or faking a foreigner’s speech pattern doesn’t fool experts for a minute. “People write the way they would talk to a foreigner, namely using all verbs in the infinitive form, while getting much more complicated elements right,” elaborates Karin Pittner. The Darknet, where the IP addresses of website operators cannot be tracked, keeps experts busy, too. Here, the subjects of linguistic analyses include drug and human trafficking platforms – the amount of texts is staggering.
A linguistic fingerprint?
If nothing is known about the author of a document, linguists are still able to identify their social and educational background, and often even their native language and gender. Is there a linguistic fingerprint after all? “The existence of such a thing is highly controversial,” says Karin Pittner. “It is more of a pointer,” believes Steffen Hessler. Be that as it may, forensic linguistics is booming, because it is a useful tool for exposing criminals. Analysis the language is also a helpful tool when it comes to automatic identification of fake news, hate speech or fake ratings at online portals.
Human gaze is indispensable
“Based on a specific style, it is frequently possible to assess if a post was authored by radical left-wing or radical right-wing groups,” says Karin Pittner. The human gaze is nevertheless indispensable, because the technology is still very much in its infancy.
Further information, as well as other interesting topics about IT-Security, you’ll find at the RUB Newsportal.
General note: In case of using gender-assigning attributes we include all those who consider themselves in this gender regardless of their own biological sex.