ID of anonymous posters via word analysis?

All the talk recently about multiple fake identities reminded me
of a research project I read about a few years ago.   A team set
out to ascertain if Shakespeare was really one person, or actually
several people.

The researchers analyzed the frequency distribution of words
found in the works of Shakespeare, and compared them to the other
writers of the day.     I don't recall the results of the
project, but that kind of research would have implications for
anonymous postings.

It is not too difficult to see how certain spelling errors, word
frequency (how often do you say 'I':-) choice of wording, and the
working vocabulary of an individual could  allow you to
identify an anonymous poster.  This would be particularly easy if the
individual also posted under their real name.

I suspect that the government has done research on this topic. It
would be useful to identify which terrorist made which (written)
threat.

This brings up the subject of how one can post without
leaving an "ASCII fingerprint".  I suspect the use of a spelling
checker and grammatical checker would help.    Perhaps running
your text through a language converter, (say English to French)
then back would remove many identifying characteristics.



Jim Pinson                     Galapagos Islands
PGP key available by finger    jpinson@fcdarwin.org.ec

I remember reading some time ago that the Nazis had a method of 
trapping communists that went like this:
  They'd put the suspect in  room and carry on a conversation with him 
(all males, as I remember). Then, after several hours of conversation 
about any and all kinds of subjects, they would tote up the number of 
times that the suspect used the word "concrete" -- evidently a favorite 
left-wing buzzword in the '30s. If the "concrete" cropped up frequently 
.... off to camp went the suspect.
  I think that identification by buzzwords, habitual misspellings, etc. 
could be used to identify anonymous posters. Sentence structure is also 
revealing. Le style, c'est l'homme, said Voltaire.  Of course, it all 
comes down to how much time and effort you want to put into proving, say, 
that SBoxx=LDetweiler.
  And the method is fallible. I've seen some pretty wacky schemes 
"proving," by word or grammatical analysis, that Shakespeare's works were 
written by Francis Bacon, Chris Marlowe, Queen Elizabeth, etc. It's very 
easy to be misled by your desire to prove something you're already 
convinced of for other reasons.