> In his book on Hume, Deleuze says this definition is nonsense. And I
> agree. Under this definition, everyone is an empiricist. Or no one is.
> It doesn't capture what empiricism really investigates, which is
> really relation between things and their movement, instead of the
> nature of things. If you accept that definition, Marx is certainly an
> empiricist.
Yeah... I guess there's a big risk of putting up straw men here and talking past each other using different definitions. What I meant by 'empiricism' in economics was the idea that there was 'too much theory' and we need to stick to 'the facts'. Maybe this is a straw empiricism, but similar things have been argued within (and against) economics with varying degrees of sophistication. My view is that any inductive argument from empirical regularities is liable to break down when the structures generating the observable variables change. Theory is necessary because we don't have direct empirical access to the structures, we build abstract conjectures about them.
However I agree that we appeal to experience in making those conjectures, and if our theories turn out to be incompatible with the observable data, there's something wrong with the theories. If that's what people mean by empiricism, then I'm an empiricist too. Some people might say that much economics is too rationalist and not empirical enough - in building great mathematical systems from axiological starting points (enshrining methodological individualism, etc) - and there's some truth in that, but I would argue that that's bad theory (if it's assumed to say something about reality) rather than too much theory.
Cheers, Mike