This article attempts to construct a mid-level theoretical overview of communication evidence (sometimes referred to as ‘forensic linguistics’), using decision trees as a theoretical framework. It suggests that there are three main issues – the nature of the sample, the type of evidence and the analysis used. The types of evidence are categorised into language crimes, tradenames, meaning transfer and author attribution, and examples and discussion are given for each. They are further subcategorised, and distinguishing features and characteristics are discussed. It is suggested that the type of
analysis used to provide linguistic evidence needs to take into account simultaneously the method of analysis and the communication systems involved. Communication systems in turn are categorised according to ‘levels’ and variable features. This article is regarded as programmatic rather than conclusive – it is intended to stir debate, rather than be seen as a firm statement.