I have updated our 2006 paper on language pattern analysis to detect social network attacks. Some minor formatting changes were needed, given the last time I generated the PDF was 2011. The original post is here.

Attacks by scammers appear to make sophisticated use of language ideology to abuse trust relationships. Language that indexes a social group allows perceived “authenticity” to be constructed in a way that breaks down a victims’ defenses — a variety of linguistic devices are used as attack tools.