> Hi Tom
>
> I confess I have only scanned the papers. Rebecca has looked at them
> in more detail, but I do think the problems of this sort of analysis,
> using computer programs, fairly intractable, and the more especially
> when talking of highly conscious imaginative verbal activity like
> writing poetry...
You'd hope so, wouldn't you? In actual fact, statistical analysis of
patterns of word usage, grammatical idom and so on has been shown to exert a
disconcerting degree of traction when it comes to things like determining
the gender of an author. I can imagine a sort of reverse Turing test, where
you have to fool the computer into thinking you're a different gender, or
not suicidal, or whatever - an interesting exercise...
Dominic
|