<< Yes, that crude empiricism again. David Laibman told me, after
> > hearing my criticisms of efficient market theory, that you can't
> > refute a theory with empirical observations, only with another
> > theory. I think that's silly, but I guess I'm just a crude empiricist.
>
>
> At the risk of being condemned as silly myself, lwhile I don't know if
> Laibman said this, I think it is a perfectly sensible thing to say.
It depends on what you think a theory should do and what the goals of
theory should be. If you're a realist (I am) and truth matters, then
theory should try and be a literally true description of the causal
structure of the world. What else is a theory supposed to do?
Well, I agree.
> Empirical
disconfirmations and puzzles then become important. While I don't think
a single disconfirmation is enough to falsify a theory, a number of
unexplained puzzles should make the theorist begin to look at the rival
theories.
And Kuhn would agree.
> There's always the inference to the best explanation which is
important in economics and politics where no theory, except Marxism,
has any explanatory or predictive power. Marxism has its problems but it
is still the best theory compared to the others. Marxism is also the
best bet for explanatory unification too, it explains the most amount of
empirical phenomena with the minimum amount of hypothesis.
Yes, and?
> One of the problems in neo-classical economics is that it has no
confirming instances, so the the practitioner of this theory must
explain away *all of empirical reality*. I think even the most adamant
non-Popperian would say that when the empirical counterexamples badly
outnumber confirming instances, the theory should be junked whatever the
alternatives are.
I think this is too glib. NCE seems to me to a lot of pretty good work empirically in the short to medium rabnge,a nd ir picks out tendencies tahtr help expalinw hy market economiesa re as dynaminic as they are in a number of ways. As to the most adamant neo-Popperian, I don't know this fellow, but his clsoe relative, Imre Lakatos, would not say that. Like Kuhn, he would say (in hsi own idiom) that when a research program dengerates, you need a better one beforee you can leave it.
I> n the case of NC economics, empirical reality itself is unimportant and
illusory. Friedman himself said that truth value is unimportant, its
predictions that matter.
That's F, but his instrumentalism is not necessareily widely shared or necessary to the theory.
> However, if the truth of the assumptions is not
important, the theory will not even approximate reality.
That remains to be shown. There area lot of false theories that do a good job of approximating reality. Classical mechanics. Webesian sociology. General relativity. Marxism.
> Then what is
the point of the theory? (besides a legitimation of capitalist
ideology.)
But look, I don't want to be backed into the position of defending NCE,a lthough I obviosult like ita lot more tahn you do. The point about philosophy of science I was making is valid, or anyway no less valid, even if you are right and NCE is wholly voodoo.
> If I remember, one of Kuhn's ideas was that the empirical evidence
used in counterexamples is dependent on the theory you are testing.
That is true, both of Kuhn and of science.
> So
theory comparison and epistemic justification generally become a social,
intersubjective thing rather than just a straigtforward comparison with
the 'world as it actually is'. I think things start getting circular
here.
Well, yes and no, Kuhn makes this quick leap, but I think it is too quick, However, if there is circularity, it will not be resolved bys cientific realist tub thumping. (Yes, the entities our theory posits are really, _really_ REALLY _REAL_.) Rilly!)
--jks