Examines errors in the language used in a particular corpus and attempts to identify patterns of mistakes.