ID of anonymous posters via word analysis?

[gtoal@an-teallach.com (Graham Toal)]

In article <9310292305.AA22193@anon.penet.fi> an41418@anon.penet.fi writes:
 > If studying writing patterns is a viable method for
 > discovering the true identity of a nym, then by all means
 > try it. If someone were to discover my identity, I would

Much can be hidden by some mechanical process of canonizing ones
text, but I'm fairly sure a lot of personal traits will remain; especially
ones to which the writer is unconscious, such as the use of particular
bits of vocabulary.  Take this pretentious word 'nym', for instance,
that some of the cypherpunks are so fond of.  Do you realise only *FOUR*
non-anonymous people on the whole of this list have used it in the last
month or so? - Tim May, Jamie Dinkelacker, Hal Finney and Richard 
Kennaway.  (There was also a remailed anonymous poster, and the cryptically
named major@dcd.wa.gov.au who interestingly posts from a .au address but
uses US English; and whose finger demon says 'no such user')

If one were slightly paranoid in the Detweiler mould, one might look
more closely at an41418's postings and the postings of the above, for
other similarities.  For instance, only Tim May and Jamie Dinkelacker
have ever used the word "essay"... (one I chose to look up at random);
or one could watch for idiosyncracies such as a misspelling like
'revokation' (Richard Kennaway alone, consistently), or "somesuch"
(Tim May, consistently) which they're probably not aware of and would
be hard pushed to suppress if they were adopting other personae.

Other things to look at are punctuation - do they consistently put
a comma before the last element of a list preceding an 'and', or do
they miss it out (as in the old style).  Things like that are a lot
harder to consciously suppress than using...ellipses for instance---instead
of dashes, in an attempt to disguise your most obvious writing traits.

Just FYI, I don't think any of the people mentioned here are each other,
unless they do a much better job of hiding their identities than I have
a right to expect - for instance Jamie Linkelackers appalling solecisms
- "thou doth protesteth too much" and "is purposive to" seem too archetypal
to me to be deliberate mistakes, and none of the other posters come close
to that kind of mistake. 

G
-- 
Personal mail to gtoal@gtoal.com (I read it in the evenings)
Business mail to gtoal@an-teallach.com (Be careful with the spelling!)
Faxes to An Teallach Limited: +44 31 662 4678  Voice: +44 31 668 1550 x212