A part of this blog referred to detecting linguistic anomalies and profiling based on the contents.
This is an update blog entry: Concordia University Anonymous Email Profiling
Researchers from Concordia University, have now achieved anonymous email detection with around 80-90% accuracy to identify the author of the email. But, the requirement for such analysis requires access to past email conversations of the suspects, this method would prove beneficial only when complete data-set is made available.
Real-time on the fly analysis using this technique, for an email system would require immense data storage and computational resources. So as of this moment the only ones to benefit from this are the governments and law enforcement agencies. Applications are few but as of this moment this seems outrageous to my mind – detection of content written by known entities using aliases with possible impact on National Security.