[OT] Text analysis and author, anyone done it?

[Peter Alcibiades]

My own purpose is not to do with religious texts, but it seems like the
example is getting in the way of the subject.  There's no reason why
statistical analysis shouldn't work on textual analysis.  Whether it does,
and which method is best, and which language is best to do it in, these are
all empirical questions to be settled by studies.  I haven't had time to do
much work on the problem yet, but it seems at first glance as if LC is a
reasonable choice.  Its going to be a tradeoff between speed and statistics
libraries ready made, which you would get in R, and the string manipulation
and rapid development that you get with LC.

I'm quite skeptical about the idea that when we write for different
purposes, we vary the parameters which are used for textual analysis.  One
of the key characteristics of a native speaker is the ability to create an
infinite number of grammatically correct different expressions.  This
doesn't happen randomly, its rule based, and its a bit like the gait when
walking, it is a quite deep feature of our use of language. Its a reasonable
hypothesis that even when writing for different purposes and audiences, our
writing will be generated in accordance with the same rules.

But its not a matter for belief, this is an empirical question.  One could
take, for instance Bertrand Russell, who wrote for specialist and lay
readers.  There are novelists - Graham Greene wrote what he called
'entertainments' as well as his serious novels.  Some UK ministers have
written polemical pieces and also fiction. Academics often write for learned
journals and also popularize.  So it would be quite easy to test.

First however, I shall try to master the basic tools.

--
View this message in context: http://runtime-revolution.278305.n4.nabble.com/Re-Text-analysis-and-author-anyone-done-it-tp3636729p3641646.html
Sent from the Revolution - User mailing list archive at Nabble.com.