> Philip Pilkington wrote:
>
> > they begin with some attempt to impose their framework
> > onto some piece of empirical data
>
> As do you, as does everyone. Empirical data never speaks for itself.
>
Obviously, but I'm saying that the framework should, as far as possible (and the possibilities are quite large), self-generate out of the data itself... however, it was the latter steps that I was really objecting to, or at least warning against: construct model, purge model of offending elements.
Even with chaos style theories you'll see this to a certain extent - they study broad patterns after all - but they'll certainly be not only more accurate, and also less biased. If a bourgeois economist begins his theoretical "inquiries" it will, as we all know, start with specifically bourgeois notions. If they were forced to adopt a framework based on allowing the facts to speak for themselves (i.e. to naturally allow patterns to manifest themselves, begin with a blank slate rather than the germ of a conceptual model) then it would be far more difficult to insert a bias.
I believe that the similarities between this approach and Marx's are striking. Note how he allows his conceptual framework to gradually form out of the observations themselves rather than imposing them on the material. Marx is constantly contradicting himself and thus allowing patterns to form of their own accord. Its all about giving primacy to the object over the subject (or in Marxese: giving primacy to man in general rather than certain men in particular); that's what Adorno called non-identity thinking.