Modeling human behaviour with higher order logic: Insider threats