On Sun, 9 Dec 2001, Hakki Alacakaptan wrote:
> The challenge is to define what makes marxism a useful diagnostic/prognostic
> tool. How do we define the effectiveness of a fuzzy-logic social theory like
> marxism without relying on concepts like falsifiability (how can you falsify
> something that's not precisely defined?)? Maybe we could assess the hit rate
> of predictions made by marxists. How about putting up a vote page with
> questions like: What will the US jobless rate be one year from now? How many
> US citizens will be tried for political crimes or treason by the end of
> 2002? How many countries will the U.S. bomb in 2002? How many new U.S.
> military bases will there be by 2004 in countries where U.S. military
> operations took place during Enduring Freedom?
>
> Hakki
These really aren't fair comparisons, are they? If you try to use theoretical principles in physics to predict behavior of a particular leaf falling on a particular fall day, your "hit rate" will be quite low. To say that probabilistic theories are too fuzzy, can't predict specific events, etc, is beside the point. I know philosophers of science have spilled a lot of ink about this (and I expect jks to come swooping in any moment here), but scientific theory is just an explanation: it explicates the relationship among variables, and the explanation generates predictions that can be empirically verified. All the other stuff--precision, parsimony, elegance-- help us to assess the quality of the scientific theory, but they are not intrinsic to scientific theory.
For instance, I'd ask NC this. Social psychologists for about 40 years have asserted that exposure to aggressive mass media content provokes aggressive behavior in people. The basic theoretical assumption here is simple: people tend to imitate the behavior of people they have observed in similar contexts. Granted, it's no general theory of relativity, but researchers have conducted lab studies, longitudinal surveys, and archival research over the past forty years--hundreds of scientific studies!--that clearly support this simple theory. In what sense is this not a well-verified scientific theory?
If you say, but this doesn't allow us to predict the behavior of a single individual who plays violent video games, then we're back to the leaf example: theories in physics or chemistry don't provide flawless predictions of individual entities in complex environments either!
I agree that generating and testing scientific theories in the human sciences is challenging, but it's far from impossible.
Miles