Tim Grant talking about a fascinating (though really spooky) field of language studies: forensic linguistics.
It is no news in computer-mediated communication research that a close linguistic analysis of computer-mediated discourse can reveal a lot about the identity of the communicator, but I find it mind-blowing how we went from the early '90s "on the Internet, nobody knows you're a dog" to being able to tell whether the person is male or female ( see for instance Asta's and Susan Herrings paper) to actually being able to provide evidence in court about the authorship of text messages in a murder case.
I am not entirely sure about the idea of the Internet users becoming monitored and fully re-traceable - after all it somehow stands against the whole concept of the freedom of speech, prejudice-free communication and decentralised information the net has always represented. On the other hand, however, I find it reassuring to know that Tim Grant and his colleagues are out there, analysing and mapping for instance the language use of teenage girls, so that they can train middle-aged police officers to impersonate girls in the hunt for pedophiles.
It was also quite fascinating to find out that my observations about the linguistic and discursive strategies managers use in a virtual team to "do" or "communicate" their hierarchical position were similar to the communicative strategies of gang leaders during the riots. One of these strategies, breaking your message at unexpected points (in linguistic terms non-transition relevant places) is a way to make sure no-one interrupts you in real-time chat and you can go on talking.
What this means, essentially, is that now even the frequency or timing of hitting 'enter' or 'send' can be a tell-tale sign of who you are and who you want to come across as in a group.
I told you it was spooky.