Forensic linguistics is the application of knowledge from the field of Linguistics to legal or forensic problems. One of these problems is the analysis of documents of disputed authorship to recognise or identify the writer. These documents could be texts such as threatening letters but also crucial piece of evidence in a criminal case, such as a set of text messages that are incriminating and that the alleged writer denies having written despite them being found on their phone. This task of identifying a writer purely from language used is traditionally called authorship analysis or authorship attribution.

This talk will firstly explain how this problem is currently solved using real casework as examples. The rest of the talk will then cover the reason why these methods work so effectively. The argument being made is that these methods capture a representation of the distinctive and potentially unique grammar of a person. The logical consequence of this fact is that there is no reason why these techniques can only be applied to writing. Authorship analysis is therefore not really about authors because the same techniques can also complement the analysis speech. The conclusion of the talk is that our own personal grammar is a behavioural biometric that can be added to the toolbox of forensic scientists for the analysis of both written and spoken language.